FE Exam Preparation Book VOL1 OSG202
Vol.1
FE Exam
Preparation Book
Preparation Book for Fundamental Information Technology Engineer Examination
Part1: Preparation for Morning Exam
Part2: Trial Exam Set
INFORMATION-TECHNOLOGY PROMOTION AGENCY, JAPAN
FE Exam Preparation Book
Vol. 1
Table of Contents
Part 1
PREPARATION FOR MORNING EXAM
Chapter 1
Computer Science Fundamentals
1.1 Basic Theory of Information
1.1.1 Radix Conversion
1.1.2 Numerical Representations
1.1.3 Non-Numerical Representations
1.1.4 Operations and Accuracy
Quiz
1.2 Information and Logic
1.2.1 Logical Operations
1.2.2 BNF
1.2.3 Reverse Polish Notation
Quiz
1.3 Data Structures
1.3.1 Arrays
1.3.2 Lists
1.3.3 Stacks
1.3.4 Queues (Waiting lists)
1.3.5 Trees
1.3.6 Hash
Quiz
1.4 Algorithms
1.4.1 Search Algorithms
1.4.2 Sorting Algorithms
1.4.3 String Search Algorithms
1.4.4 Graph Algorithms
Quiz
Questions and Answers
i
2
3
3
7
10
11
14
15
15
18
21
24
25
25
27
29
30
32
34
37
38
38
41
45
48
50
51
Chapter 2
Computer Systems
62
2.1 Hardware
2.1.1 Information Elements (Memory)
2.1.2 Processor Architecture
2.1.3 Memory Architecture
2.1.4 Magnetic Tape Units
2.1.5 Hard Disks
2.1.6 Terms Related to Performance/ RAID
2.1.7 Auxiliary Storage / Input and Output Units
2.1.8 Input and Output Interfaces
Quiz
2.2 Operating Systems
2.2.1 Configuration and Objectives of OS
2.2.2 Job Management
2.2.3 Task Management
2.2.4 Data Management and File Organization
2.2.5 Memory Management
Quiz
2.3 System Configuration Technology
2.3.1 Client Server Systems
2.3.2 System Configurations
2.3.3 Centralized Processing and Distributed Processing
2.3.4 Classification by Processing Mode
Quiz
2.4 Performance and Reliability of Systems
2.4.1 Performance Indexes
2.4.2 Reliability
2.4.3 Availability
Quiz
2.5 System Applications
2.5.1 Network Applications
2.5.2 Database Applications
2.5.3 Multimedia Systems
Quiz
Questions and Answers
ii
63
63
65
68
70
73
77
79
81
83
85
85
87
89
90
95
99
100
100
102
104
106
108
109
109
111
113
116
118
118
121
123
125
126
Chapter 3
System Development
138
3.1 Methods of System Development
3.1.1 Programming Languages
3.1.2 Program Structures and Subroutines
3.1.3 Language Processors
3.1.4 Development Environments and Software Packages
3.1.5 Development Methods
3.1.6 Requirement Analysis Methods
3.1.7 Software Quality Management
Quiz
3.2 Tasks of System Development Processes
3.2.1 External Design
3.2.2 Internal Design
3.2.3 Software Design Methods
3.2.4 Module Partitioning Criteria
3.2.5 Programming
3.2.6 Types and Procedures of Tests
3.2.7 Test Techniques
Quiz
Questions and Answers
139
139
141
143
144
147
149
151
154
155
155
157
159
162
163
165
167
170
172
Chapter 4
Network Technology
181
4.1 Protocols and Transmission Control
4.1.1 Network Architectures
4.1.2 Transmission Control
Quiz
4.2 Transmission Technology
4.2.1 Error Control
4.2.2 Synchronization Control
4.2.3 Multiplexing and Communications
4.2.4 Switching
Quiz
4.3 Networks
4.3.1 LANs
4.3.2 The Internet
4.3.3 Various Communication Units
4.3.4 Telecommunications Services
Quiz
Questions and Answers
iii
182
182
184
187
188
188
190
192
194
195
196
196
198
200
202
204
205
Chapter 5
Database Technology
212
5.1 Data Models
5.1.1 3-layer Schemata
5.1.2 Logical Data Models
5.1.3 E-R Model and E-R Diagrams
5.1.4 Normalization and Reference Constraints
5.1.5 Data Manipulation in Relational Database
Quiz
5.2 Database Languages
5.2.1 DDL and DML
5.2.2 SQL
Quiz
5.3 Control of Databases
5.3.1 Database Control Functions
5.3.2 Distributed Databases
Quiz
Questions and Answers
213
213
215
217
218
221
223
224
224
226
231
232
232
234
236
237
Chapter 6
Security and Standardization
244
6.1 Security
245
245
6.1.1 Security Protection
6.1.2 Computer Viruses
247
6.1.3 Computer Crime
249
Quiz
251
252
6.2 Standardization
6.2.1 Standardization Organizations and Standardization of Development and
Environment
252
6.2.2 Standardization of Data
254
6.2.3 Standardization of Data Exchange and Software
256
Quiz
258
259
Questions and Answers
iv
Chapter 7
Computerization and Management
7.1 Information Strategies
7.1.1 Management Control
7.1.2 Computerization Strategies
Quiz
7.2 Corporate Accounting
7.2.1 Financial Accounting
7.2.2 Management Accounting
Quiz
7.3 Management Engineering
7.3.1 IE
7.3.2 Schedule Control (OR)
7.3.3 Linear Programming
7.3.4 Inventory Control (OR)
7.3.5 Probability and Statistics
Quiz
7.4 Use of Information Systems
7.4.1 Engineering Systems
7.4.2 Business Systems
Quiz
Questions and Answers
Part 2
262
263
263
265
267
268
268
270
274
275
275
278
282
284
286
290
291
291
293
296
297
TRIAL EXAM SET
Trial Exam Set
309
Fundamental IT Engineer Examination(Morning) Trial
Answers and Comments
Fundamental IT Engineer Examination(Afternoon) Trial
Answers and Comments
v
310
339
381
419
2
Computer Systems
Chapter Objectives
A computer system is composed of hardware and software. There are many types of computer, but the principles of their operation are fundamentally the same.
We will learn the mechanism of computers (hardware) in
Section 1 and software (operating system) for efficient computer use in Section 2. We will further learn some configurations of computer systems for achieving improved reliability in Section 3 and ways to evaluate the performance of computers in Section 4. Finally, in Section
5, we will learn various systems that use computers.
2.1
2.2
2.3
2.4
2.5
Hardware
Operating System
System Configuration Technology
Performance and Reliability of Systems
Systems Application
[Terms and Concepts to Understand]
Central Processing Unit (CPU), cache memory, input/output interface, auxiliary memory, task management, job management, multiprogramming, virtual memory, dual system, duplex system, client/server system, availability, MTBF, MTTR, Internet
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
62
2. Computer Systems
2.1 Hardware
Introduction
Hardware is defined as devices with which a computer is configured. Computers consist of processing units, memory, input/output units, etc. In this section, we will explain these units from the standpoint of hardware.
2.1.1 Information Elements (Memory)
Points
¾
¾
Information elements include ROM and RAM.
SRAM and DRAM are typical types of RAM.
Semiconductor memory is memory made of integrated circuits (ICs) using semiconductors.
Semiconductor memory includes ROM, which is non-rewritable, and RAM, which is rewritable. ROM (Read-Only Memory)
ROM is semiconductor memory that is not erased when the power is turned off.1 On mask
ROM and PROM, data can be written only once; on EPROM, however, data can be written and re-written repeatedly using a special method. Types and characteristics of ROM are shown in the following table.
Type (Name)
Mask ROM
PROM (Programmable ROM)
EPROM (Erasable PROM)
EEPROM (Electrically EPROM)
Flash memory2
Characteristics, etc.
Data is written at the time of manufacturing. It cannot be re-written later.
PROM data is written by the user when it is first used. It cannot be re-written later.
EPROM data is written by the user electrically. All the data can be erased using ultraviolet rays.
All the data can be erased and re-written. Data is erased electrically. Erasure and re-writing can be done collectively or on a block basis. Data is erased electrically.
1
Volatility and non-volatility: It is the property that the contents of memory are lost when the power is turned off is called volatility. RAM is a type of volatile memory. On the other hand, the property that the contents of memory are not lost when the power is turned off is non-volatility. ROM is a type of non-volatile memory.
2
(Hints & Tips) Flash memory is classified as EEPROM.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
63
2. Computer Systems
RAM (Random Access Memory)
RAM is semiconductor memory that loses its memory contents when the power is turned off.
Unlike ROM, its contents can be changed, so it is used for main memory, graphics memory,3 and cache memory.
There are two typical types of RAM: SRAM and DRAM. The characteristics of SRAM and
DRAM can be summarized as shown in the following table.
Comparison item
Level of integration
Access speed
Price
Usage
Operation
Structure
SRAM
Low (small capacity)
Fast
Expensive
Cache memory4
Battery-operated devices
No refresh is required.
Flip-flop
Complicated structure
DRAM
High (large capacity)
Slow
Inexpensive
Main memory
Refresh is required.
Condensers and transistors
Simple structure
SRAM (Static RAM)
SRAM is composed of a flip-flop,5 so it does not require any refresh operations and is able to speed up information reading and writing. However, the cost is higher for the same capacity than DRAM, because the SRAM structure is more complicated than that of DRAM. For this reason, it is used mainly in areas where the speed, not the cost, is important, such as in cache memory. It is also used in battery-operated devices.
DRAM (Dynamic RAM)
DRAM consists of condensers and transistors, representing whether or not there is electrical charge in the condensers, by using 1 or 0. As time elapses, the electrical charge in the condensers gets discharged, resulting in memory loss; therefore, it needs to be re-written
(refreshed) at certain time intervals (every few milliseconds). Since the structure is rather simple, the manufacturing cost is low, and it is mainly used in the main memory of PCs.6
Types of DRAM equipped with high-speed data transfer functions include SDRAM, DDR
SDRAM, etc.
3
Graphics memory: It is memory used when images and characters are displayed on the display screen using a computer.
It is also referred to as video memory (VRAM).
4
Cache memory: It is high-speed memory placed between main memory and the CPU to speed up data reading from main memory to the CPU.
5
Flip-flop (also know as bistable circuit): It is an electrical circuit with two stable states, which maintains its state until an input that changes one of the states is entered.
6
(FAQ) Frequently there are questions that compare SRAM and DRAM. You should have good knowledge of the level of integration, usage, structure, etc.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
64
2. Computer Systems
2.1.2 Processor Architecture
Points
¾
¾
A computer consists of five major units (functions).
There are several addressing methods; direct addressing, indirect addressing, etc.
The term architecture refers to “structures or organizations.” The processor architecture refers to the configuration and operating principles of the computer.
Configuration of Computer
Below is a figure showing the basic configuration of a computer. This configuration is called the “big five units” or “big five functions,”7 because there are five major components.
Processing unit
Control unit
Control flow
Data flow
Operation unit
Input unit
Memory
Main memory Output unit
Auxiliary memory The control unit and the operation unit are together called the processing unit or the central processing unit (CPU).
Address Modification and Addressing Methods
A program is stored in the main memory and is retrieved, one instruction at a time, by the control unit to be executed. Address modification occurs in order to locate the location of data that is subject to processing.
Address modification is a function that obtains the value of the address actually accessed based on the address specified by the instruction. The method for address modification is called an addressing method. The address actually accessed as the result of the address modification is called the effective address. Addressing methods is described below in detail.
7
•
•
•
•
•
“The big five units”:
Control unit: It is the unit that controls the entire computer. It extracts and reads instructions of the program stored in the main memory and sends to various units the directions necessary to execute the instruction.
Operation unit: It is the unit that performs the arithmetic operations, logic operations, and other operations. It consists of adders, registers, complementers (units that convert values to their complements), etc.
Memory: It is a generic term of the unit that stores data, programs, etc. It can be classified into main memory and auxiliary memory.
Input unit: It is a generic term of the unit that enters programs and data into the computer.
Output unit: It is a generic term of the unit that outputs results of computer processing in characters and numbers that we can recognize.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
65
2. Computer Systems
Direct addressing method
In this method, the content stored in the address part of the instruction becomes the data subject to operation.
（Direct addressing method）
Instruction
part
Address part x Address
Main memory
∼
0
1
2 x
Indirect addressing method
In this method, the data stored in the address designated by the address part of the instruction are not the data subject to operation; rather, the data stored at the address designated by that content are the data subject to operation.
（Indirect addressing method）
Instruction
part
Address part x Address
Main memory
0
1
x
y
y
Indexed addressing method (index modification)
In this method, the effective address is the sum of the value of the address of the instruction and the value of the index register.8 For example, when processing an array, we can look up the content of another address simply by changing the content of an index register.9
（Indexed addressing method）
Instruction
part
i
Address part x Address
Index
Index register
0
1
0
1
x
y
8
Main memory y
x+y
Register: It is low-capacity, high-speed memory where data is temporarily stored. It is located in the CPU. There are various registers, including the following: general registers, for storing intermediate and final results of operations; status registers for indicating the CPU state after an instruction is executed; index registers for address calculations; and base registers. 9
(FAQ) There are questions on the concept of addressing methods. An example is “Which of the following is an appropriate description of the direct addressing method?” Be sure to have these methods organized in your mind: the direct addressing method, index addressing method, immediate value addressing method, etc.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
66
2. Computer Systems
Base addressing method
In this method, the effective address is the sum of the address designated by the address part of the instruction and the content of the base address register.10
（Base addressing method）
Instruction
part
B
Address part x Address
Index
Base register
0
1
B
0
1
x
b
Main memory y
b+y
Immediate value addressing method
This is the method where the address of the instruction stores the data subject to processing, not an address.11
RISC and CISC
Various types of technology have been used to speed up computer processing time; RISC and
CISC are examples of this technology. With the advancement of semiconductor technology, the integration density of integrated circuits has risen continually. At present, computers are made with integrated circuits, and RISC and CISC are the two approaches for developing computers.
Computers configured for high speed with simple instructions and simplified hardware are called RISCs. In contrast, those where complicated instructions are configured on one circuit are called CISCs.
RISC (Reduced Instruction Set Computer)
These computers have only a set of simple, frequently used instructions integrated onto a single
VLSI (very large scale integration) chip in order to achieve high performance through improved machine cycles (operation speed) and a reduction in instruction processing time. The emphasis is placed on keeping the length of each instruction to a fixed length and limiting the time required to execute each instruction to a fixed amount. By doing so, the technology of pipeline control has been easily implemented. However, the number of instructions to be executed becomes large unless efficient object programs are created, so it is essential that the compiler have an optimization function.12 Most computers called workstations are of this type.
CISC (Complex Instruction Set Computer)
These computers have complex instructions integrated onto a single VLSI ship in order to achieve high overall performance. Most general-purpose computers are CISCs.
10
(Note) The base addressing method can be used regardless of where in the main memory the program is stored, simply by changing the value of the base register. Such a structure is called a re-locatable structure.
11
(Hints & Tips) Note that in the immediate value addressing method, the address of the main memory is not designated.
12
Optimization: It is a function of a compiler to eliminate redundancy of a program in order to reduce the execution time of the object program and the size of the program. This is done in a variety of ways, such as calculating constants in advance, simplifying formulas, and eliminating double loops.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
67
2. Computer Systems
Pipeline Control
We have mentioned RISC and CISC as technologies to improve computer processing speed. To further improve the speed, the RISC system uses pipeline control. Pipeline control is a technology to reduce the instruction execution time of the CPU. This is an attempt to do the following: when execution steps of an instruction are divided into 5 or 6 steps, and if each step is completed within a certain fixed amount of time and the instruction steps stay independent of one another, then we can improve the overall processing speed by delaying the execution of each instruction 1 step behind the previous instruction. In reality, however, due to branching instructions, there are times when the next instruction address is not completely determined, and some steps are not completed within the fixed processing time. Everything is not always functioning in an ideal way, but pipeline control does process instructions concurrently, providing one way of speeding up the computer.13
Execution order
Time
Instruction 1
I1
Instruction 2
I2
I3
I4
I5
I1
I2
I3
I4
I5
I1
I2
I3
I4
I5
I1
I2
I3
I4
Instruction 3
Instruction 4
I1 : Fetch an instruction
I2 : Decode
I3 : Calculate an address
I4 : Fetch data
I5 : Execute
I5
2.1.3 Memory Architecture
¾
Points
¾
A hierarchical memory structure is introduced in order to achieve high-speed, large-capacity memory.
Cache and interleaving technologies are used for speeding up memory. There are many requirements for memory, but requirements for high speed and large capacity14 are of particular importance. In general, however, high-speed memory is expensive and has small capacity whereas low-speed memory is inexpensive and has large capacity. So, efforts are being made to combine high-speed but small-capacity memory and low-speed but large-capacity memory to develop high-speed and large-capacity memory.
13
(FAQ) Questions regarding pipeline control often appear on exams. Most of them are in the form of choosing an appropriate description of pipeline control, so you only have to know that pipeline control executes instructions concurrently. 14
(Note) Besides these, requirements for memory include reliability, ability of random access, non-volatility, re-writable function, portability, low cost, etc.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
68
2. Computer Systems
Memory Hierarchy
Memory hierarchy is a hierarchical representation of the relationship between the access speed and capacity of various types of memory.15
Expensive
price
speed
Fast
Inexpensive
Slow
Register
Cache memory16
Main memory
Disk cache
Auxiliary memory (hard disk, etc.)
Large-capacity memory (optical disk, etc.)
Memory capacity
Cache Memory (High-Speed Buffer Memory)
Cache memory is high-speed, small-capacity memory that is placed between the CPU or registers and the main memory. The main memory is slower than the CPU or registers, so the
CPU process can be made more efficient by storing frequently accessed data and programs of the main memory into the cache memory. Recently, a secondary cache has been installed for further improvements in speed.
The cache memory placed between a hard disk and the main memory is called the disk cache.
The figure below shows the relationship between the cache memory and the disk cache.
Central Processing Unit (CPU)
Cache memory
Primary cache
Secondary cache17
Main memory
Disk cache
Hard disk unit
15
(Note) If t is the average memory access time, tm is the access time of the main memory, tc is the access time of the cache memory, and h is the hit ratio, then the following equation holds: t = tch + tm(1 – h).
16
Hit ratio: It is the probability that the portion of a program necessary to execute that program is in the cache memory
17
Secondary cache: The primary cache is the cache memory which is built in the CPU; the secondary cache is the cache memory placed between the primary cache and the main memory.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
69
2. Computer Systems
Interleaving (Memory Interleaving)
The main memory is divided into multiple units called banks, and addresses are assigned across the banks. Often, the main memory is accessed over a sequential range of addresses at a time, so the speed can be enhanced by accessing a sequential range of addresses concurrently. For example, even when the operation of Bank 0 is not yet completed, Bank 1 can be accessed.
Below is an example of 4-way interleaving (with 4 banks).
Bank 0
0
4
Address
Address
Bank 1
1
5
Bank 2
2
6
Bank 3
3
7
…
…
…
…
Central Processing Unit (CPU)18
Data and programs are stored over a sequence of addresses (horizontally), but the memory is accessed in bank units (vertically). This allows concurrent access to a sequence of addresses.
2.1.4 Magnetic Tape Units
Points
¾
¾
Capacity calculations require blocking factors and record density.
Performance calculations do not include the stop time.
Magnetic tape is a medium that records data onto a tape that has been magnetically coated.
The unit price of this memory medium is cheap and has a large capacity, so it is used in cases such as backing up entire hard disks.
Capacity Calculation
The record format of a magnetic tape is shown below. As we can see in this figure, in order to record one block, we needs to include an IBG (Inter-block gap) which is an area to identify the block and contains a special code. A magnetic tape reader reads data in block units, so this area is crucial even in identifying the end of each block.
Record
length
…
IBG
Record 1
Record 2
Record 3
…
Record n
IBG
…
Block length
Tape length necessary to record one block
18
(FAQ) Interleaving is a way to speed up memory. Questions on the concept of interleaving have appeared often, so be sure you understand this.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
70
2. Computer Systems
Assuming that the specifications of a magnetic tape are given below, let us calculate precisely the number of records that can be stored on this single magnetic tape.19
[File specifications]
Record length
Blocking factor
[Magnetic tape specifications]
Record density
Inter-block gap (IBG)
Tape length
80 bytes
100
64 bytes/mm
15 mm
730 m
Calculation of block length
The blocking factor is 100. Æ 100 records can be stored in one block.
Record length Æ 80 bytes
So the number of bytes L1 for each block, excluding the IBG, is as follows:
L1 = 80 (bytes/record) * 100 (records/block) = 8,000 (bytes/block).
The record density is 64 bytes/mm. Æ 64 bytes can be stored on 1 mm of tape.
Hence, the length L2 of a block, excluding the IBG, is as follows:
L2 = 8,000 (bytes/block)/64 (bytes/mm) = 125 (mm/block).
Therefore, the block length L3, including the IBG, is as follows:
L3 = 125 (mm/block) +15 (mm/block) = 140 (mm/block).
80 bytes20
…
IBG
Record 1
Record 2
15mm
Record 3
…
Record n
IBG
…
L1=8,000 bytes, L2=125mm
L3=140mm
Number of records that can be stored on one magnetic tape
Let us now calculate precisely how many records can be stored on one magnetic tape.
(1) Calculating the number of blocks that can be stored on one tape
Since the length of the tape is 730 m (730 * 103 mm), the number B1 of blocks that can be recorded on one tape is as follows:
B1 = (Length of the tape) / (Length of a block)
= 730 * 103 / 140
19
(FAQ) On each exam, there is at least one question dealing with the calculation of the capacity or performance of a magnetic tape or a hard disk. If you keep these ideas organized in your mind, you can answer these questions because the difference is only in the numerical values.
20
(Hints & Tips) Besides “bytes/mm,” the record density can be represented in “columns per mm” or “bpi.” A column is the same as a byte. “bpi” stands for “bytes per inch,” which is the number of columns per inch. Converting from inches to mm is necessary here, but you need not remember the formula since the relationship between inches and mm will be given in the question.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
71
2. Computer Systems
= 5,214.285…≒5,214 (blocks).21
The fractional part 0.285 is less than one block, so we truncate it.
(2) Calculating the number of records that can be stored on one tape
Since one tape can record 5,214 blocks, and each block has 100 records, the number B2 of records that can be stored on one tape is as follows:
B2 = (number of blocks that can be stored on one tape) * (blocking factor)
= 5,214 * 100
= 521,400 (records)
Performance Calculation
The running speed of the tape is constant when reading or writing data. In theory, the tape begins to accelerate in the middle of IBG and starts to read and write when a constant speed is achieved. When another IBG is found, the tape decelerates and stops in the middle of IBG.
IBG
block
IBG
block
IBG
Running speed of the tape
Time
Stop
Start
Running at a constant speed
Constant Deceleration speed Assuming that the specifications of a magnetic tape are given below, let us calculate the time it takes the magnetic tape to read one block.22
[Specifications of a magnetic tape]
Data transfer speed
Record length
Blocking factor
Start or stop time
320 Kbytes/sec
80 bytes
100
6 milliseconds
Transfer time for one block
Transfer time for one block can be obtained by dividing the length of a block by the data transfer speed:
Data transfer time for one block = (block length) / (data transfer speed)
Block length = (record length) * (blocking factor)
= 80 * 100
= 8,000 (bytes)
Data transfer time for one block = 8,000 (bytes) / 320 (Kbytes/sec)
21
(Hints & Tips) The fractional part of the number of blocks is discarded here, but it actually becomes a short block, which is a block with fewer records than the other blocks.
22
(Hints & Tips) In performance calculations, data is transferred in blocks, so we do not need to consider the length of
IBG.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
72
2. Computer Systems
= 8,000 / (320 * 103)
= 25 * 10-3 (seconds) = 25 (milliseconds)
Time required to read one block
We add the start-up time to the data transfer time for one block.
Time required to read one block
= (start-up time) + (data transfer time for one block)
= 6 (milliseconds) + 25 (milliseconds) = 31 (milliseconds)
Note that the start-up time is added but not the stop time. When being read, no data is transferred until the beginning of a block is reached. Hence, the time until this is achieved is part of the waiting time. After that, data is transferred, but when the data transfer is completed, the stop operation and the program processing are performed concurrently. Thus there is no need to add the stop time.23
2.1.5 Hard Disks
Points
¾
¾
A hard disk is configured with cylinders and tracks.
Hard disks of the sector type do not have IBGs.
A hard disk is a medium that achieves random and high-speed reading and writing of data, consisting of 1 to 10 round disks coated with a magnetic substance on the front and the back sides and rotated at a high speed. If there is only one disk, it is called a floppy disk.24
High-speed rotation
Some disks do not record data on the very top and very bottom sides (protective sides).
Tracks
(Circumference portion)
Read/write head
(portion that reads and writes data)
(as many as the recording sides of the disk)
Cylinders
(Cylindrical portion)
Access arm
23
(Hints & Tips) That stop time is not included is a standard assumption in exam questions.
(Note) Another medium that, like a floppy disk, can read and write and can easily be carried around is MO (magneto optical disk), which is very popular because its capacity is about 600 times that of a floppy disk.
24
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
73
2. Computer Systems
Capacity Calculation
Just as on a magnetic tape, data is recorded in blocks on a hard disk. However, if a block cannot fit into a track, it cannot be recorded.
Block 1
IBG
Block 2
IBG
…
IBG
block n
Block
Cannot be recorded
Track length
First, data is recorded on a track, and when that track is filled, the recording proceeds to the next track which is a track on the corresponding circumference of the next surface. In other words, data is recorded in cylinder units.
Assuming that the specifications of a hard disk are given below, let us calculate actually how many cylinders are necessary to write 100,000 records.
[File specifications]
Record length
Blocking factor
250 bytes
8
[Specifications of a hard disk]
Number of cylinders per disk
Number of tracks per cylinder
Number of bytes per track
Block gap
400
19
13,000
135 bytes
Block length (B)
B = (record length) * (blocking factor) + (block gap)
= 250 * 8 + 135 = 2,135 (bytes)
Number of blocks that can be recorded on one track (N)
N = track length / 2,135 = 6.008…≒6 (blocks) (truncated)
We truncate the answer because a fractional part does not constitute a block.
Number of records that can be recorded on one track (Rt)
Rt
= N * (blocking factor)
= 6 * 8 = 48 (records)
Number of records that can be recorded on one cylinder (Rs)
Rs
= Rt * (number of tracks per cylinder)
= 48 * 19 = 912 (records)
Number of cylinders required to write 100,000 records (S)
S
= (number of records) / Rs
= 100,000 / 912 = 109.649… ≒110 (cylinders) (rounded up).
In general, files are secured in cylinder units, so if there are unused tracks on a cylinder, the
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
74
2. Computer Systems
number of cylinders is rounded up to the next integer.25
Performance Calculation
The access time of a hard disk is calculated as follows:
Access time = waiting time + data transfer time
= (seek time + latency time) + data transfer time
Translator’s note: The waiting time (seek time + latency time) is often called the access time.
The seek time is the time during which the read/write head moves to the track where the data is recorded. The latency time is the time until the desired data come under the read/write head.
The seek time and the latency time are determined by where the head is located, so we use the average values. In actual exam questions, the seek time is always given, and the latency delay can be calculated by the duration of one rotation, which is obtained by the inverse of the number of rotations per time unit. Since the minimum latency time is 0 and the maximum is 1 rotation time, the average latency time is the duration of a half rotation.
Assuming that the specifications of a hard disk are given below, let us calculate actually the access time for reading data contained in one block (5,000 bytes).
[Specifications of a hard disk]26
Number of rotations of the hard disk:
Memory capacity per track:
Average seek time:
2,500 rotations per minute
20,000 bytes
25 milliseconds
Calculation of the average latency time
The fact that the number of rotations of this hard disk is 2,500 rotations per minute means that the disk makes 2,500 rotations every minute. Hence, the rotation time is as follows:
Rotation time = (1 * 60,000 msec/min) / 2,500 revolutions/min = 24 msec/revolution
Note carefully these units. The rotation speed is given in rotations per minute, but the rotation time is in milliseconds. Hence, we need to convert minutes to milliseconds (1 minute = 60 seconds = 60,000 milliseconds).
Since the average latency time is ½ of the rotation time, it is 12 milliseconds.
Calculation of data transfer speed
Since one rotation allows the transfer of data contained on one track, 20,000 bytes are transferred in 24 milliseconds. Hence, 20,000 / 24 (bytes/msec) is the data transfer speed. We can calculate this quotient. But, since it is indivisible, we shall leave it as it is here.
25
(FAQ) On each exam, there is at least one question dealing with the calculation of the capacity or performance of a magnetic tape or a hard disk. If you keep these ideas organized in your mind, you can answer these questions because the difference is only in the numerical values.
26
Seek time/Latency time: The seek time is sometimes called the positioning time. The latency time is sometimes called the search time.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
75
2. Computer Systems
Calculation of data transfer time
The time required to transfer 5,000 bytes is then calculated as follows:
Data transfer time = (amount of data transfer) / (data transfer speed)
= 5,000 / (20,000 / 24)
= 5,000 / 20,000 * 24
= 6 (msec)
Calculation of access time
Access time = 25 msec + 12 msec + 6 msec = 43 msec
The calculations above are based on the rotation speed of the hard disk. However, auxiliary memory, such as a hard disk and a magnetic tape, exchanges data with the computer through input/output channels. 27 It is therefore necessary to install input/output channels with appropriate transfer speeds.28
Capacity Calculation of Sector-based Hard Disk
The term sector refers to the way a magnetic medium is partitioned on floppy disks or hard disks. A sector is an arc of a fan-shaped portion of the disk formed by radial lines drawn from the center of a track in equal intervals. Input/output of a sector-based recording medium is done in sector units, without using IBGs. Each sector is filled with as many records as possible, and the remaining portion of the sector is not used.
Track (on circumference)
Sector (arc)
For example, suppose that each track consists of 12 sectors, each of which consists of 1,200 bytes on a hard disk. To store files whose record length is 900 bytes, there is no sector remainder, as shown below, if the flocking factor is 4.
1,200
3,600
1,200
1,200
Sector
Sector
Sector
Record
Record
900
900
Record
Record
900
900
← 3 sectors are used
← 4 records in a block
3,600
27
Input/output channel: Data-transfer path for exchanging data between auxiliary memory and the computer
(Hints & Tips) The data transfer speed of a hard disk is determined by the rotational speed of the disk, so it is meaningless to have a high-speed channel. Channels are to be selected according to the transfer speed of the disk.
28
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
76
2. Computer Systems
2.1.6 Terms Related to Performance/ RAID
¾
Points
¾
Regarding storage media, there are some terms such as access time, waiting time, transfer time, seek time, and latency time.
RAID is a disk array system for achieving enhanced reliability and/or increased processing speed.
There are some terms related to the performance of storage media, including access time, waiting time, transfer time, seek time, and latency time. RAID, sometimes called a disk array, is a way to control multiple disks placed in parallel as if they were one unit.
Terms Related to Performance
The figure below shows how the various processing times are related to one another, beginning at time S, when a processing unit requests input/output to a hard disk unit, to time T, when the data delivery is completed.
Access time
Waiting time
Transfer time
Seek time
Latency time
S
I/O request
T
Delivery complete
Since the head moves to the desired track while the disk is rotating, the rotation time and moving time partially overlap.29 However, on IT exams, this overlap is almost always ignored.
Hence, there is no problem defining waiting time as follows:
Waiting time = Seek time + Latency time (or Search time)
RAID (Redundant Array of Independent Disks)
RAID describes auxiliary storage in which multiple hard disks are placed in parallel and are controlled as if they were one disk unit so that the input/output speed can be improved and/or reliability can be enhanced. Sometimes the term RAID refers to such an auxiliary storage or a method. It is an attempt to speed up the process by spreading the blocks over multiple disks and reading the blocks simultaneously.
29
(Hints & Tips) A hard disk is rotating as the head approaches the track, so the seek time and the latency time overlap partially. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
77
2. Computer Systems
Main memory
←Block
←Block
←Block
←Block
Disk 1
Disk 2
Disk 3
Disk 4
RAID has 6 levels as described below, from RAID0 to RAID5.
RAID0
This is the method of writing blocks of a fixed size on multiple disks. Access is not centralized on one single unit, so the input/output time can be reduced.30
RAID1
By recording the same data on two disks, this configuration enhances the safety of the data.31
RAID2, RAID3, and RAID4
These are configurations where, in addition to data recorded on the hard disk, there is a disk designated as the error-checking disk to prevent failures. RAID 2 can correct errors. RAID3 and RAID4 can detect errors while they cannot correct errors. In RAID3, data is partitioned in bits or bytes whereas RAID4 partitions data in block units.32
RAID5
This is where each data block is assigned a parity value. Data and parity are written on separate disks, and a failure on a single disk can be recoverable.
Below is an example of RAID5. Here, 4 disks are handled as one group.
Disk 1
Block 1
Block 4
Block 7
Parity 10~12
Disk 2
Block 2
Block 5
Parity 7~9
Block 10
Disk 3
Block 3
Parity 4~6
Block 8
Block 11
Disk 4
Parity33 1~3
Block 6
Block 9
Block 12
Here, the data is divided up into blocks of a certain length, and three blocks are considered to form a unit. For example, Blocks 1 through 3 are a unit, and for each bit, the exclusive logical sum of blocks 1 through 3 is written on a separate disk as parity value 1 through 3. Similarly, the exclusive logical sum of blocks 4 through 6 is written on a separate disk as parity value 4 through 6.
There is also RAID6, in which the parity values are separated as in RAID5 and the data is recoverable even when two disks fail.34
30
(Hints & Tips) The only feature about RAID0 is that I/O is dispersed, so this is not a measure to improve reliability.
(Note) RAID1 is called mirroring since the same data is recorded on separate hard disks.
32
(Note) There is also RAID0+1, a combination of RAID0 and RAID1, already in use.
33
(Note) The parity of RAID5 uses the exclusive logical sum of multiple blocks. Hence, even if one of the disks should fail, the damaged data can be recovered by taking the exclusive logical sum of the other blocks.
34
(FAQ) Often on the exam, there are questions of the form: “Which of the following statements is appropriate concerning
RAID…?” Remember the difference between RAID0 and RAID1.
31
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
78
2. Computer Systems
2.1.7 Auxiliary Storage/Input and Output Units
¾
Points
¾
Auxiliary storage includes hard disks, magnetic tapes, magneto optical disks, CDs, DVDs, etc.
Input/output units include keyboards, image scanners, tablets, displays, printers, etc.
Any storage excluding the main memory is called auxiliary storage. Auxiliary storage can be used to compensate for the insufficient capacity of the main memory. In general, auxiliary storage has larger capacities in comparison with the main memory.
Input and output units include both input units, where data is entered into the computer, and output units, where data is taken out of the computer. A unit equipped with both the input and output functions is called an input/output unit.
Auxiliary Storage
Typical auxiliary storage includes the following media. In the past there was a time when magnetic tapes and floppy disks were the mainstream media; however, recently the main types have been hard disks, magnetic optical disks, CDs, and DVDs.35
Medium
Hard disk
DVD
DVD-ROM
DVD-R
DVD+R
DVD-RAM
DVD-RW
DVD+RW
CD
CD-ROM
CD-R
CD-RW
Capacity small to 300GB
4.7 to 9.4BG
3.95 to 7.9GB
Re-writing
Yes
No
Yes, only once
3.95 to 7.9GB
Yes, many times
700MB
700MB
700MB
No
Yes, only once
Yes
Magneto optical disk36
(MO)
Floppy disk
Hard disk
DAT37
128, 230,
640MB, 1.3GB
1.4MB
a few GB
Max. about
24GB
Yes
Yes
Yes
Yes
35
Properties, etc.
Mostly built-in
Replacement for CD-ROM
Playable on DVD-ROM units Multiple specifications
For software distribution, etc.
For backup, etc.
Requires a dedicated drive for re-writing
Written with magnetism and light and read with light
Good portability
For backup
For backup
(Note) DVDs are optical disks just as CDs are, but with reduced laser-light wavelength, the DVDs have larger capacities.
The record density on DVD is also larger.
36
(Hints & Tips) A magneto optical disk uses light and magnetism for writing data but uses only light for reading data.
37
(Note) DAT is a unit that records audio onto a magnetic tape using digital signals. It was originally designed for music, but it is now used as a backup system because of its low cost.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
79
2. Computer Systems
Input and Output Units
Input units include keyboards, image scanners, tablets, pointing devices, etc. Output units include displays, printers, etc.
Input units
The most common input devices are keyboards and pointing devices. A keyboard is used to enter numerals and characters while pointing devices38 are used to enter coordinate values.
Other input units are shown in the following table.
Unit
OCR
OMR
Mouse
Tablet
Barcode reader Image scanner
Functions, etc.
This reads handwritten characters and printed characters optically.
This reads handwritten marks optically.
This is used to enter coordinate positions of the mouse pointer.
Dedicated devices such as a light pen are used to enter coordinate positions. This reads barcodes (not the numbers printed).
This reads image data such as pictures and photos, converting them to digital data.
Output units
The most common output devices are displays and printers.39
Unit
Descriptions, etc.
This uses a cathode-ray tube; it is inexpensive and
Display CRT comes with a large screen.
This uses liquid crystal; it is expensive but thin and
Liquid crystal40 saves footprint.
The principle is the same as for a copy machine; the
Printer
Laser printer print quality is good, but it is expensive.
Printing is by injection of ink; it is small and
Inkjet printer inexpensive. The printing quality is good but requires specific paper.
Thermal printer
Heat melts the ink, resulting in high-quality, but the
Thermal-transfer
operating cost is high. printer This is noisy but inexpensive; this has duplicating
Dot impact printer capability. Printer for design drawings
Plotter
38
(Note) Pointing device: A pointing device is any unit designed to enter coordinate positions such as a mouse, a tablet, etc. Other examples include trackballs, digitizers, touch screens, etc.
39
(Note) OLED: It is a display using the organic light emitting diode technology. It uses organic materials that emit light when voltage is applied. Compared with LCDs, the viewing angle is larger, the contrast is better, and the response speed is higher, in addition to being thinner and lighter.
40
(Note) Liquid crystal display (LCD): A liquid crystal display can be of various types such as TFT and STN (currently the mainstream is DSTN). STN has a simple structure with low manufacturing costs, but its resolution and contrast are also low. DSTN is an improved version of STN, where contrast is enhanced. TFT has contrast and resolution equivalent to those of CRT but is expensive.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
80
2. Computer Systems
2.1.8 Input and Output Interfaces
Points
¾
¾
I/O interfaces include SCSI, USB, etc.
USB is equipped with a hot plug function and a plug-and-play function. Input and output interfaces are interfaces for connecting peripheral devices such as printers and hard disk units to the PC and for transferring data. Depending on their types, the transfer may be either serial data transfer or parallel data transfer.41
Data Transfer Methods
0
1
0
1
0
1
0
1
Peripheral devices
Computer
There are two data transfer methods between the computer and its peripheral devices: serial transfer and parallel transfer. Serial transfer is the type of transfer in which data output from the computer are serially transferred, one bit at a time. Parallel transfer is the type of transfer in which data bits are transferred in parallel from the computer. For example, if the transfer is
8-bit parallel, there are 8 signal lines:
Peripheral devices
Computer
Serial transfer
Parallel transfer
41
(FAQ) Frequently the exams have questions concerning combinations of I/O interfaces and data transfer methods. It is good to know common I/O interfaces and data transfer methods.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
81
2. Computer Systems
Types of Input and Output Interfaces
Below is a table that summarizes I/O interfaces commonly used today. USB is the most commonly used interface now, equipped with a plug-and-play function42 and a hot plug function.43 Type
RS-232C
Transfer method
Serial
SCSI
Parallel
Centronics
Parallel
GPIB
Parallel
USB
IEEE1394
44
Serial
Serial
Properties, etc.
Originally used to connect a PC to a modem; currently used also for I/O units
Daisy chain allows connection of up to 7 relatively high speed units
Connecting a PC to a printer; high speed but not for long distance Connecting a PC to a measuring device; also known as
IEEE-488
Connecting up to 127 units in a tree configuration
Connecting up to 63 units in daisy chain or tree configuration Daisy chain
This is the connection method used in SCSI and GPIB, where peripheral units are connected along a line. The last unit in the line requires a termination resistor called a terminator.
External hard disk
MO drive
CD drive
Terminator
(to be placed at the end)
USB
USB has two modes: the full speed mode of 12Mbps and the low speed mode of 1.5Mbps. In the full speed mode, relatively high speed units such as printers and scanners are connected. In the low speed mode, relatively slow units such as keyboards and mice are connected. Currently,
USB 2.0 has increased its high speed mode up to 480Mbps, so most peripheral units can be connected. Other interfaces
In addition to the above, there are other interfaces such as IDE (connecting a hard disk), ATA
(IDE standardized by ANSI), and ATAPI (connecting ATA with units other than a hard disk, such as CD-ROM drive and tape streamer). However, currently CD-R and CD-RW units are commonly connected via SCSI and USB.
42
Plug-and-play (Plug and play): This refers to the function of automatically installing and setting the device driver when the peripheral unit or extension card is connected to the computer. The OS checks all the units connected to the computer when it is started up, installing the required device drivers. If the OS does not have the device driver of the unit in its own library, it requests installation of the device driver and, if necessary, even re-starts the computer automatically.
43
Hot plug: This is the function that enables the plug-and-play function while the computer and peripheral unit power is on. 44
IEEE 1394: The standard where the transfer speed is 100Mbps, 200Mbps, or 400Mbps. This is equipped with a hot swap function (peripheral units can be connected or disconnected without having to turn the power off).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
82
2. Computer Systems
Quiz
Q1
Compare DRAM and SRAM:
Item for comparison
Refresh
Level of integration
Access speed
Unit price per bit
Usage
Q2
DRAM
SRAM
Among the components that compose a computer, which ones compose the central processing unit (CPU)?
Control unit
Control flow
Data flow
Operation unit
Input unit
Memory
Main memory
Output unit
Auxiliary memory Q3
Explain the role of cache memory.
Q4
Describe the characteristics of RAID.
Q5
Is USB a serial interface or a parallel interface?
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
83
2. Computer Systems
A1
Item for comparison
Refresh
Level of integration
Access speed
Unit price per bit
Usage
DRAM
Required
High (large capacity)
Slow
Inexpensive
Main memory
SRAM
Not required
Low (small capacity)
Fast
Expensive
Cache memory
A2
Collective term for both control unit and operation unit
A3
It is high-speed small-capacity memory placed between the CPU or the register and the main memory. The main memory is slower than the CPU or the register, so the
CPU processes can be made more efficient by storing frequently accessed data and programs of the main memory in the cache memory.
A4
This is auxiliary memory in which multiple hard disks are placed in parallel and are controlled as if they were one disk unit so that the input/output speed can be improved and/or reliability can be enhanced. Sometimes the term RAID refers to such a method instead of the storage. It is an attempt to speed up the process by spreading records over multiple disks and reading the distributed records simultaneously.
A5
Serial interface (“USB” stands for “Universal Serial Bus.”)
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
84
2. Computer Systems
2.2 Operating Systems
Introduction
Computers do not operate only with hardware. They function only with the use of software called an operating system (OS).
2.2.1 Configuration and Objectives of OS
Points
¾
¾
There is a broadly defined OS and more narrowly defined OS.
The objective of an OS is the effective use of computers.
The definition of an operating system (OS) is not clear. The basic software is called an OS in a broad sense while the control program is called an OS in a narrow sense.
Configuration of OS
An operating system is the basic software that comprehensively controls and manages the entire operation of hardware and software of a computer system. A program referred to as the basic software and its role are shown below. 45
Basic software
Control program (OS in a narrow sense)
- Job management - Data management - Operation management
- Communications management - Task management
- Memory management - Fault management - Miscellaneous
Service program
- Linkage editor - Sort/merge program - System generator - Text editor
- Media conversion program - Debugger -Miscellaneous
Language processor
- Compiler
- Assembler
- Interpreter
- Generator
- Miscellaneous
Service programs and language processors are sometimes called processing programs, which run on the control program. For this reason, the control program is called an OS in a narrow sense.
45
(Note) Operating systems for personal computers include Windows XP, Mac OS X, and OS/2. Those for workstations include Windows Server 2003, UNIX, and Mac OS X Server. For general-purpose machines, there is also MVS developed by IBM. In addition, there is a free OS program called Linux, which is compatible with UNIX.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
85
2. Computer Systems
Objectives of OS
Objectives of OS
An OS attempts to improve the productivity of the entire system by eliminating unnecessary operations and waste of various resources surrounding the computer and by operating the computer system efficiently. The objectives of an OS are organized in the following figure.
Effective use of hardware resources:
Multiprogramming,46 spooling function,47 etc.
Response to various processing modes:
Batch processing, online real-time processing, etc.
Securing reliability and safety:
Improvement in RASIS, etc.
Load reduction of application software:
Virtual memory,48 library management,49 etc.
Support of computer control and operation:
Continuous processing, recording the operation data, etc.
Let us take a look at these individual objectives in detail:
Effective use of hardware resources
Hardware resources include the central processing unit (CPU), memory, I/O units (including channels). etc. The OS controls these resources so that they can be used efficiently.
Response to various processing modes
One computer can handle various processing modes such as batch processing, remote batch processing, online processing, real-time processing, and interactive mode processing. In particular, since online processing has become widespread, the scope of computer applications has been dramatically enlarged.
Securing reliability and safety
Indexes for reliability and safety include RASIS. This is a term coined by taking the initial letters of the words Reliability, Availability, Serviceability, Integrity, and Security.
Load reduction of application software
Application software refers to a program which runs under the control of the OS. The OS provides an environment in which application programs can be efficiently executed.
46
Multiprogramming: It is a mechanism in which programs are processed alternatingly on one CPU so that it can appear as though multiple programs were operating at the same time.
47
Spooling: It is accomplished by using high-speed hard disks as a virtual I/O unit. For example, directly printing on a low-speed printer slows down the processing speed. Instead, the output results can be recorded on a high-speed hard disk first, and then a service program, dedicated only to output, can do the printing when the CPU is not busy.
48
Virtual memory: It is a technique to enlarge the apparent capacity of the main memory so that large-scale programs can be loaded in the memory at a time. Often, auxiliary storage such as a hard disk is used as virtual memory.
49
Library management: It is the function that systematically accumulates primitive programs, object programs, load modules, and other programs developed. This enables integrated management of software assets that are managed individually (by individuals).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
86
2. Computer Systems
Support of computer control and operation
An OS eliminates human intervention as much as possible, as it processes programs (jobs) continuously and records the operation status (log). The record of the operation status is used to check the circumstances under which a fault occurred.50
Human
intervention
Job 1 preparation
Idle time
OS
Resources free
Job 1 execution
Job 1 preparation
Job 2 preparation
Job 1 end
Job 1 execution
Job 2 execution
Job 1 end
Job 2 preparation
Resources free
Automated
Idle time
Job 2 execution
Job 2 end
2.2.2 Job Management
¾
Points
¾
Jobs are units of tasks given to the computer, consisting of multiple programs (job steps).
Job management has functions such as scheduling and spooling.
One of the functions of the control program, the “OS" in a narrow sense, is "job management.” In job management, the priorities of jobs are determined, and the jobs are synchronized. In batch processing, the OS analyzes the contents of JCL (job control language) to assign resources51 and schedule jobs. In interactive mode processing, the OS analyzes instructions entered at the terminal, assigns resources, and performs scheduling. In addition, job management has other functions such as spooling and cataloged procedures.
Job management
Scheduler:
Managing the order of job execution
Master scheduler:
Interface with the operator
Job scheduler:
Managing reception, selection, start, and finish of jobs
Reader:
Reading jobs
Initiator:
Preparing for the beginning of jobs and programs
Terminator:
Spooling:
Clean-up after jobs and programs
Input management for jobs, output management for process results
Cataloged procedure:
Support for execution of typical jobs
50
(FAQ) Concerning operating systems, many exam questions involve knowledge of terms. Be sure to know terms such as multiprogramming, virtual memory, and spooling function.
51
Resource: A resource is a device/unit of various kinds necessary for the computer to operate. It refers to any device related to memory, input, output, control, and other functions; specifically, these include the CPU, main memory, and files.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
87
2. Computer Systems
Scheduler
In job management, jobs are continuously executed under a master scheduler and job scheduler.
The master scheduler plays the role of an interface with the operator via the console panel.52 The job scheduler manages the reception, selection, start, and finish of the jobs.
Reader
This reads the contents of JCL, analyzes them, schedules jobs, and places them in a queue.
Initiator
This selects the programs with high execution priorities among those in the queue and assigns the resources that those programs need.
Terminator
This releases resources that were used by programs just completed. If there is another program following, the terminator starts up the initiator.
Spooling (Spool)
Spooling is the function of the I/O of jobs independent of the programs. Any output results to low-speed units such as a printer are first stored in a spool file. Then, after the program is finished, the output results are printed on the printer from the spool file by the service program of the OS.53
The reason this is done is that, when the I/O unit is slow, directly performing the I/O process would reduce the processing speed of the computer.54
Cataloged Procedures
In job execution directions, typical processing (routine work) such as translation of languages is done in the following way. A set of JCLs is registered together at a separate location, and this registered set of JCLs is called for executing programs. By doing this, the computer prevents JCL errors. This set of registered JCLs is called cataloged procedures.
52
Console panel: It is a unit where the operator interacts with the computer system via key control and monitors its operation and where the system communicates failures, etc. It consists of a keyboard and a display.
53
(Note) For spool files usually a hard disk is used. The service program that performs the sending of spool files to a printer is often called a writer (output writer).
54
(FAQ) Most exam questions on job management are about spooling. They are always in the form of selecting the correct term, so be sure to have accurate understanding of spooling.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
88
2. Computer Systems
2.2.3 Task Management
Points
¾
¾
A task is the smallest execution unit for using a resource.
An interruption takes place in order to switch tasks.
One of the functions of the control program, which is the “OS” in a narrow sense, is “task management.” Task management is the function of controlling the execution of programs and consists of various procedures such as synchronization control of programs, dynamic assignment of resources for program execution, and management of execution priorities of the programs. It also conducts various types of interruption control.
Tasks and Jobs
A unit of processing from the perspective of the user (a single program or a set of programs executed consecutively) is called a job. In contrast, a unit of internal processing executed under the OS is called a task. “Task” refers to a processing unit that subdivides a program process.55
Control of Task Execution
Task management generates tasks required in response to a command and monitors the execution process. When a task generated becomes no longer necessary, the task is eliminated. The state transitions for tasks are shown below:
Running
2. Dispatching
3. Timer interruption Task generation 1. Registration
Ready
5. Input/output interruption 6. Terminate
Task elimination 4. Supervisor call
Waiting
1. A task has been generated, so it is entered into the queue. Æ Move to the ready state.
2. For execution, move to the running state via the task dispatcher. Æ Move to the running state.
3. Time has expired; withdrawn to make way for a task with high priority. Æ Move to the ready state.
4. Withdraw for an I/O instruction. Æ Move to the waiting state.
5. Again waiting to be executed after completion of I/O. Æ Move to the ready state.
6. All processes are now completed. Æ The task is terminated.
Dispatching refers to the step of selecting a task with high priority from among those tasks in the ready state and advancing it to the running state.56 Supervisor call refers to the step of invoking a function of the OS; in state transfer of tasks, this refers to an I/O instruction. I/O interruption is a
55
Process: This term refers to a program being executed and is used interchangeably with the term “task.” “Process” is a word used by some operating systems such as UNIX. In recent years the expression “process” is used frequently.
56
Dispatcher: It is the program of the OS that carries out dispatching; also known as the dispatching routine.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
89
2. Computer Systems
notice that I/O is completed.57
Interruption
Interruption refers to temporary suspension of a program currently being executed for any reason and transferring control to the OS to execute some necessary processing program. There are external interruptions caused by certain specific states of the hardware and internal interruptions caused intentionally when the control program is called from within a program.
The hardware detects interruptions. When the CPU detects an interruption, the OS receives it, changes the program being executed to the necessary state prior to the interruption, examines the cause of the interruption, and transfers control to the corresponding processing routine (program).
The area in which this state is stored is called the PSW (program status word). There is also a possibility that while an interruption process is taking place, another interruption occurs. Priorities are given to these interruptions depending on their types so that multiple interruptions can also be controlled. The table below describes main types and examples of causes for interruptions.
Type of interruption cause Name
Machine check interruption
External interruptions
Clock function interruption
(timer interruption)
Input/output interruption58
Internal interruptions 59
(traps)
External signal interruption
Program interruption
Supervisor calls (instruction interruption) Possible causes
Malfunction of units, fault, power/voltage trouble
Elapsed time of a fixed length (interval timer), reaching a designated time
Input/output completion, input/output unit status change (out of paper, etc.)
Instruction from console panel, external signals
Overflow, underflow, undefined instruction code execution, division by 0, memory protection violation
Input/output operation command, task switching, page fault, control program invoked
2.2.4 Data Management and File Organization
Points
¾
¾
Data management provides integrated methods for accessing files.
File organization includes sequential organization, direct organization, indexed organization, partitioned organization, etc.
Another function of the control program, which is the “OS” in a narrow sense, is “data management.” Data management is the control program that manages data input and output. It provides various file organization methods such as sequential organization, direct organization, and indexed organization. It works as a bridge between logical files processed within a program and physical files whose structures are different.
Data management allows programmers not to worry about the physical structure of the files.
57
(FAQ) State transition of tasks (processes) is almost certain to appear on every exam. Commit the entire figure of state transition to your memory.
58
(Hints & Tips) Issuing of an I/O instruction is notified by supervisor call; I/O completion is notified by I/O interruption.
59
Internal interruptions are intentionally caused by programs, so they are sometimes referred to as traps.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
90
2. Computer Systems
Access Methods
Access methods
Access methods include sequential access, direct access, and dynamic access, which is a combination of the first two.
Sequential access:
Direct access:
Dynamic access:
Processing files sequentially from the beginning60
Processing a specific record directly
Using direct access to find and position a record, followed by sequential access
Sequential access
This is the method of handling records in a file in sequential order from the beginning. This can be performed with almost all recording media. This is suitable for collective processing in which all records in a file are subject to processing.
Direct access
This is the method where a necessary record is directly (randomly) accessed regardless of the order in which the records are stored. This method is used when the file medium is a directly accessible storage medium such as a hard disk. It is used in online real-time systems where only a part of a large number of records stored in a file needs to be accessed for quick update.
Dynamic access
This is the method where direct access is used to find a specific record and then sequential access follows. Similar to direct access, this is used when the file medium is a directly accessible storage medium.
60
(Hints & Tips) Sequential access can be used with most media; however, direct access and dynamic access are limited to directly accessible media such as a hard disk.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
91
2. Computer Systems
File Organization
File organization
File organization methods include sequential organization, direct organization, indexed organization, and partitioned organization,61 etc.
Sequential organization files
Files are processed in the order in which they were recorded on the medium.
Direct organization files
Records are directly processed by key values.
Indexed organization files
Direct access by index and sequential access are possible. Partitioned organization files
One file contains multiple sequential organization files. Sequential organization files
The records in the files are stored in consecutive positions following a certain order. Files of this type can be created on almost all media such as magnetic tape, hard disk, and floppy disk.
In general, only sequential access is possible with these files.
Direct organization files
A storage address on the medium is calculated based on the key value found in each record, and the record is stored in that position.62 To access a record, we first calculate the storage address using the same formula, and the record is read from that location. There is a method where the key value of each record is directly used as the storage address for the record, but this is not very practical, creating a lot of wasted memory if the key values are not consecutive. A more general way is to use a certain type of conversion formula to calculate a storage address from the key value of each record. This is called address conversion (randomization).
Key
Store
Address calculation
Record
Record
Record
Record
Record
Record
…
…
…
Record
Record
Record
Address conversion sometimes produces the same address for different records. In such a case, the record stored first is called the home record while the record assigned to the same address later is called a synonym record.63
61
(FAQ) Questions concerning the characteristics of these organization methods are frequently asked. Know the characteristics of each type of file organization.
62
(Hints & Tips) A special case of direct organization is relative organization, in which the key values are consecutive like
1, 2, 3, … and the key value is itself the storage address of the record.
63
Synonym/Home: Address conversion can take different key values but produce the same storage value. This is called a
“synonym” (word meaning the same thing). If the result of key conversion stores a record, this record is called the home record, and another record that could not be stored there is called a synonym record. A synonym record needs to be stored elsewhere by some other method (e.g. by list).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
92
2. Computer Systems
Indexed organization files
These are files with an index, and they are organized such that the user can access the records by looking up their addresses by index. Sequential access, direct access, and dynamic access are all possible. In each case, the actual records are accessed only after their addresses on the medium are looked up using the index, so not only does the medium contain a basic data area
(prime domain) where the data is stored but also an index area. Further, in order to prevent a situation where records cannot be added to the basic data area, an overflow area64 is also reserved. Partitioned organization files
In these files, sequential organization files are grouped into units called members, each of which is given a name. Then a directory containing these names and their leading addresses is created. Access is allowed to these members. Think of a member as a set of multiple files organized sequentially. Direct access can find the beginning of a member, and sequential access can be used to find a record in that member. Partitioned organization files are used as storage locations for program files and libraries but not very much as data files.
Member A
Member B
Member C
Directory area (Register)
File B
Member area
File A
64
(Note) There are two types of overflow areas: cylinder overflow area and independent overflow area. Overflow records from various tracks are stored in a cylinder overflow area; if a cylinder overflow area becomes full, additional records are stored in an independent overflow area shared by all of the files.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
93
2. Computer Systems
Hierarchical File Systems
UNIX and Windows65 make use of hierarchical file systems as mechanisms to manage files efficiently. Directories and files
A file system has a hierarchical structure consisting of files and directories (directories are file registers). At the top of the hierarchical structure is the root directory, and directories under it are called subdirectories.66
Root
DIR 1
File 0
Root directory
DIR 2
DIR 4
DIR 3
DIR 5
Subdirectory
File 1
File 2
: Directory
: File
File manipulations
When searching for a file, we designate the path showing in which directory the file is located.
There are two methods for doing this. For instance, if the hierarchical structure is as shown in the figure above, we can designate the path in the following ways:67
• Absolute path
Designating a path from the root directory68
<Example> Here is a way to designate “file1.”
\DIR3\file1 (The leading symbol “\” indicates the root directory.)
• Relative path
Designating a path from the current directory69
<Example 1> Here is the designation of file2 when the current directory is “DIR2.”
DIR4\file2
<Example 2> Here is the designation of file2 when the current directory is “DIR4.” file2 65
(Hints & Tips) What UNIX and MS-DOS call “directory” is called “folder” in Windows and MacOS.
(Hints & Tips) Whereas directories and files can be made under a directory, files and directories cannot be made under a file. 67
(FAQ) Concerning hierarchical file systems, there are exam questions like “Choose an appropriate designation as an absolute path or as a relative path.” In those questions, the symbol for separating directories and files, as well as its use, will be explained in the question text.
68
(Hints & Tips) Here we are using the symbol “\” to separate directories and files, but some operating systems use the symbol “/” instead.
69
Current directory: It is a directory in which the user is working at the moment.
66
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
94
2. Computer Systems
2.2.5 Memory Management
¾
Points
¾
Memory management uses two types of memory: real memory and virtual memory.
The basic format of the virtual memory system is the paging method.
Another role of the control program, which is the “OS” in a narrow sense, is “memory management.” Memory management makes the most effective use of the memory as well as compensating for any lack of the main memory capacity. To this end, it effectively uses auxiliary storage as part of the memory.70
Real Memory System71
This is the system that manages the physical space of the main memory. Real memory can be controlled in various ways: partition, swapping, relocation, and overlay methods.
Partitioned method
When a program is placed in the main memory, the main memory is partitioned into several partitions, into which the program is loaded.72 Without memory management, fragmentation occurs, causing a situation which prevents programs from being stored even though empty space exists. Hence, to combine all empty areas together, compaction73 is necessary.
Swapping (roll-in / roll-out)
Swapping refers to execution as the program keeps switching back and forth between the main memory and auxiliary storage. If a program is entered with higher priority than the priority level of the currently executed program, the new program is immediately loaded into the main memory and is executed. However, if there is no space in the main memory, any program in the main memory can be moved to the auxiliary storage. Hence, this system compensates for a lack of main memory capacity by utilizing auxiliary storage. However, if swapping occurs frequently, it means that programs are switched back and forth many times, thus reducing the processing efficiency of the computer system.
Main memory
Program X
Program B
Program A
Swap out (roll out)
Program X
Swap in (roll in)
Program A
70
Auxiliary storage
Memory leak: Sometimes, for some reason, memory in the main memory, secured dynamically by an application, may not get released but remains in the main memory. This is called memory leak. To eliminate memory leak, compaction must be performed.
71
Real memory system or “Real Storage (RS)”: This refers to the actually existing memory; it is the main memory.
72
(Note) In partitioned method, multiple programs can be stored simultaneously, so multitasking is possible.
73
Compaction: It means collecting empty memory areas to form a continuous area; also known as garbage collection.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
95
2. Computer Systems
Relocation (Relocatable)
Relocation refers to the function wherein a program already assigned to a certain area is re-stored in another location. A program whose structure allows it to be relocated is called a relocatable program.74 Overlay method
The physical limitations of the main memory can be eliminated; that is, programs are divided into segment units, and only the necessary segments75 are loaded into the main memory to be executed.
The entire program is stored in auxiliary storage, and the main memory contains only frequently used segments. Exclusive segments, which are never used simultaneously, are loaded from auxiliary storage to the main memory on an as-required basis.
For example, suppose that Segment A is a main routine used with high frequency while Segments
B and C are subroutines called exclusively by Segment A. While Segment B is being executed,
Segment C is in auxiliary storage. When Segment C is called, it is loaded in the area of Segment B.
Consequently, the entire memory capacity of the program is “A + B + C,” but the capacity of the main memory is sufficient if it is at least the greater of “A + B” or “A + C.”
Program
Call
Segment A
Segment B
Segment C
Load
Segment B
Segment C
Capacity actually required
Program capacity Load
(Auxiliary storage)
Segment A
(Main memory)
Virtual Memory
Virtual memory provides a large capacity of storage space regardless of the size of the main memory.76 Programs are stored in virtual memory (normally in auxiliary storage), and only the parts necessary for execution are loaded into the main memory.
Since the program is loaded into virtual memory, the instructions and data is given virtual addresses, which need to be converted to actual addresses (main memory addresses) for the execution of the program. This conversion is implemented by hardware called DAT (Dynamic
Address Translator).
If virtual memory is used, it is necessary to convert virtual memory addresses to physical addresses in the main memory. This address conversion is performed at a high speed by DAT.
Relocation is also performed if the main memory has fragmentation. Since this relocation is done during the execution of a program, it is called dynamic relocation.
We discuss the virtual memory strategies, which include three methods: page, segment, and segment-page. 74
(Note) “Relocatable” means that compaction is possible.
Segment: It is a logical processing unit of a program. Here, we can regard a segment as a subroutine.
76
Virtual memory system or “Virtual Storage (VS)”: It is a conceptual storage that does not actually exist. A program to be executed appears to be loaded into virtual memory, which is a large memory space, while only the portions (pages or segments) of the program with high frequency of use, data, and other parts necessary for the execution get loaded into the main memory.
75
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
96
2. Computer Systems
Page method (Paging method)
In this method, the program is partitioned into units of a fixed size, called pages. A page then becomes the unit for loading into the real memory. Pages are managed by a page table, which has one entry for each page of virtual memory. If the corresponding page is in the real memory, the page fault bit becomes 0. This page fault bit then indicates whether or not the corresponding page is in the real memory.
Page 0
Page n
n
n
0
b
Page 1
⋮
⋮
a n ⋮
Page table
0 a 0
1 b 0
2
1
3
1
⋮
⋮
⋮⋮
⋮
Page n
⋮
Program
Virtual memory
Page 0
Page 1
Page 2
Page 3
Page number
Real memory
Page fault bit
Dynamic relocation
Dynamic address translator
Converts to real addresses
Segment method
In this method, programs and logical sets of data is considered segments. Virtual addresses consist of segment numbers and addresses within the segments. The paging method is only for memory management and as such, the programs do not need to be written with pages in mind. In contrast, in the segment method, in which segments have different capacities, the programs must be written in consideration of the segment sizes.
Segments are logical processing units, so they can be treated as subroutines. However, the flexible lengths are sometimes inconvenient to manage, and the usage efficiency of the main memory may be reduced.
Segment-page method
This is an improved version of the segment method, in which segments are further partitioned into pages. Real addresses are accessed in the order of “segment Æ page Æ relative displacement within the page.”77
77
Relative displacement within the page: It is an address assigned such that the beginning of the page has displacement
0.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
97
2. Computer Systems
Paging Algorithms
If a page necessary for processing is not found in the real memory, an interruption called a page fault occurs, and the page is read into the real memory from virtual memory. This is called page-in. On the other hand, page-out is to move an unnecessary page out to virtual memory.
Page-in and page-out are together called paging.78
If paging occurs frequently, the time for executing the control program increases, reducing the performance. This is called slashing. To minimize the occurrence of slashing as much as possible, various algorithms are proposed to select pages that are subjects of page-outs.
Common page-out methods and their properties are shown below.
Method
LRU (Least Recently Used)
FIFO (First In First Out)
79
Properties, etc.
Pages are compared on the time elapsed since the last referencing. The page with the longest elapsed time is paged out. The page with the longest elapsed time up to the present is paged out.
78
(Hints & Tips) Swapping and paging are similar, but note that swapping takes place in program and segment units whereas paging takes place in units called pages, which are parts of a program.
79
(FAQ) There are exam questions that require specific tracking of page-ins and page-outs. For example, if the order in which pages are used is “1, 3, 2, 3, 5, 2,” and if the main memory has page 3, which page will be the first to be paged out?
For these questions, have clear understanding of ideas like LRU and FIFO.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
98
2. Computer Systems
Quiz
Q1
Explain the roles of task management.
Q2
What type of file organization is this? Sequential organization files are grouped into units called members, each of which is given a name. A directory is created, including these names and leading addresses, and access is allowed to these members.
Q3
Explain swapping.
Q4
What is the unit of loading into the main memory in the page method?
A1
It is the function that controls program execution, performing such functions as synchronized control of programs, dynamic allocation of resources to execute the programs, and managing the executing priorities of the programs. It can also conduct a variety of interruption control.
A2
Partitioned organization
A3
Swapping refers to execution as the program keeps switching back and forth between the main memory and auxiliary storage. If a program is entered with higher priority than the priority level of the currently executed program, the new program is immediately loaded into the main memory and is executed. However, if there is no space in the main memory, any program in the main memory can be moved to the auxiliary storage. Hence, this system compensates for a lack of main-memory capacity by utilizing auxiliary storage.
A4
Page
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
99
2. Computer Systems
2.3 System Configuration Technology
Introduction
Various system configurations are being used to reduce the cost and increase the efficiency of computer systems. These include client/server systems to distribute the load, dual systems to improve reliability, and duplex systems.
2.3.1 Client Server Systems
Points
¾
¾
A client server system is a typical example of distributed processing.
Clients request processing, and servers provide services (processing).
A client/server system (CSS) is a form of computer system where a network is used to distribute processing; it is a type of system configuration.80
Overview of Client Server System
CSS is configured with computers with roles called clients and servers. The servers provide services (processing) such as file management, database management, modification and supply of data, printing control, and communication functions, as requested by clients.
A client sends service requests to a server, receives the results of data processed by the server, and displays the results.
The clients and the servers distribute their processes in an attempt to spread the load of computer processing. In addition, by sharing resources, the user can reduce waste. For instance, by connecting a high-speed printer to a server, the clients can share the high-speed printer. Having one high-speed printer may be less expensive than preparing a low-speed printer for each of the clients, although this depends on the number of clients.
Computers used as a server are generally more high-performance than the client computers. To clarify the functions of various servers, they can be named by their functions, such as file servers, database servers, print servers, and communication servers.81
80
Client/server system: In multiprogramming, if there is one operational computer, the user can run both the client and the server within the one computer. In other words, a client/server process does not mean that everything is distributed. It is one method to achieve distributed processing.
81
(Hints & Tips) If a service request made by a client cannot be provided by the server, that server can become a client and request another server to perform the requested process.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
100
2. Computer Systems
Server
Process request
UNIX, Windows Server 2003, etc.
Process result
Windows XP,
Macintosh, etc.
Client
Types of Server
The table below shows the types of server, depending on the provided functions.82
Type of server
File server
Database server
Print server
Communication server
Services provided
Managing shared files, controlling file access and writing
Managing databases, operating databases with DBMS (Database
Management System)
Managing shared printer, printing when a printing request is received
Providing communication functions with the outside using a network
Characteristics of Client/Server System
Since CSS provides a typical type of distributed processing, the characteristics of distributed processing apply in a straightforward manner.83
Advantages of client/server system
•
•
•
•
•
When the processing is local at clients, the response is faster.
Costs for the entire information system can be reduced, achieving a good cost/performance ratio. Specialized server functions give the system greater economical efficiency and performance.
It is easily extended; clients and servers can be flexibly added as needed.
Even when clients access a server, they are unaware that they are in a distributed environment. Disadvantages of client/server system
•
•
•
The system could be confusing unless the server administrator is clearly identified.
The performance deteriorates if the use gets concentrated on specific servers.
The performance of the entire system depends on the network performance.84
82
(Hints & Tips) Clients and servers do not necessarily have to have the same OS. Where there are multiple servers, they do not have to have the same OS either. In addition, if there are multiple clients, they need not have the same OS either.
83
(Note) If client/server type programs are logically divided into three layers (presentation layer, function (application) layer, and data layer), such a system is called a 3-layer client server system. By distinguishing 3 layers by function, such a system strives to enhance system performance and efficiency for development and maintenance.
84
(FAQ) There have been many exam questions regarding the knowledge of client/server systems. Most of them are about the role of a client or that of a server, so be sure to know these things well.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
101
2. Computer Systems
2.3.2 System Configurations
¾
Points
¾
A dual system is a configuration for high reliability; a duplex system is a configuration for high availability.
There are two types of multiprocessor systems: loosely coupled and tightly coupled systems.
A variety of system configuration patterns are available according to the objectives of information processing. For instance, the reliability of a computer system improves by having multiple units installed. We will look at main system configurations and their characteristics from the viewpoints of reliability, efficiency, costs, and other factors.
Simplex System
This system consists of one CPU only. Reliability and the processing capabilities are inferior in comparison with other configurations, but it is economical. This configuration is commonly used.85
Dual System
This is a system configuration in which two CPUs perform the same processing and compare the processing results to each other. This configuration is applied when the process is not allowed to stop, even for a moment. If one CPU fails, the system cuts off the failed CPU and continues processing on the other CPU. Reliability is extremely high, but this system is expensive.86 87
Comparing
processing results1 85
DCE (Data Circuit-terminating Equipment): This unit converts signals received from communication line, sends them to data terminals, and also executes exactly the opposite operation. Normally, this unit is connected at the end of a communication line and functions as an interface with a computer. A modem (modulator-demodulator) is used on an analog line, and DSU (Digital Service Unit) is used on a digital line.
86
CCU (Communication control unit): This unit controls the reception and transmission of data, performs error control, and assembles and decomposes characters.
87
DISC: Auxiliary storage
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
102
2. Computer Systems
Duplex System (Standby System)
In this system, two CPUs are prepared, where the primary system is used for online processing while the secondary system is used for low-priority processing such as batch processing.88 If the primary system fails, the online processing gets switched to the secondary system.89
Switching time is said to be anywhere from several tens of minutes to several hours. Reliability is lower than that of a dual system, but this system is better from the standpoint of costs, so it is more commonly used than the dual system.
Switching unit
Switching unit
(Primary system)
(Secondary system)
Multiprocessor System (Concurrent Processors)
In this system, multiple CPUs and CCUs are sharing and processing tasks, so it is a system configuration with high processing efficiency. There are two types of multiprocessor system. One is LCMP90 (loosely coupled multiprocessors), in which multiple computer systems are controlled by separate operating systems. The other is TCMP91 (tightly coupled multiprocessors), in which multiple computer systems share the main memory and are controlled by the same operating system. The figure below shows an example of TCMP.
92
88
(FAQ) There are many exam questions on characteristics of dual, duplex, and multiprocessor systems. The key term for each system is as follows: “comparing process results” for dual, “switching the units” for duplex, and “sharing the main memory” for multiprocessor systems.
89
(Note) If a failure occurs in the primary system, it takes time to switch to the secondary system. This is because the batch processing or whatever else is being executed in the secondary system must be suspended, and the OS must be booted for the online system. A hot standby system configuration can solve this by standing by, ready to switch at any time.
In this case, the OS for the online system stays on, so switching can occur immediately.
90
LCMP (Loosely Coupled Multiprocessor): Each CPU has its own main memory and independent OS. CPUs are joined by a high-speed network or shared path. This is a configuration where independent computer systems are connected via a network.
91
TCMP (Tightly Coupled Multiprocessor): One main memory and one OS are shared in this configuration. Each CPU can perform identical processes, so even if one CPU fails, the processing can continue, albeit with lower performance. This configuration is highly reliable and thus is used in systems where a high level of processing capability is required.
92
MM: Main memory
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
103
2. Computer Systems
2.3.3 Centralized Processing and Distributed
Processing
Points
¾
¾
Centralized processing has a high level of safety but is inflexible.
Distributed processing is economical, but its real substance is hard to understand. Depending on how the computers are placed physically, there are two types of processing: centralized processing and distributed processing.
Centralized Processing
Centralized processing is a system configuration in which one computer is connected with many terminals, and the one computer alone does all of the processing. It is easy to maintain the consistency of data, and it is easy to manage the resources. These merits contributed to the popularity of this configuration in which a general-purpose computer is used as the host in centralized processing. Below is a summary of relative comparison with distributed processing.
Advantages of centralized processing
It is easy to improve the cost/performance ratio. (Grosch's Law93)
Operation and maintenance require a smaller staff. Safety level of the system is high.
Disadvantages of centralized processing
Extendability is poor to keep up with new technologies.
Backlog can be easily accumulated.94
Overhead of the OS is significant.
Recovery of a host failure is time-consuming.
A failure has a far-reaching effect.
Distributed Processing
Distributed processing is a system configuration in which multiple computers connected via a network perform the processing. Since the processing is done through a network, the processing time is longer than that of centralized processing. But, the merit is that a failure of one computer does not affect the entire system. Below is a summary of relative comparison with centralized processing.95 Advantages of distributed processing
Management responsibilities are clear. (Management responsibilities can be delegated to each organization.)
Effects of system failure are local.
Maintenance is easy (locally closed).
It is economical as only necessary units are installed
93
Disadvantages of distributed processing
Its real status can be hard to understand.
It is difficult to identify trouble spots.
Network performance has great impact.
Data inconsistency can occur easily.
Individual units are managed carelessly.
Grosch's Law: It states that “performance is directly proportional to the square of the price.” If the price of a computer doubles, the performance quadruples. However, technological advancement has reduced the prices of devices significantly, so this law is no longer applicable.
94
Backlog: It means systems, software, programs, etc. that are necessary to develop but the development of which has not even begun. The term often refers to those that are held back in the IT department within a company.
95
(FAQ) There are exam questions where you are required to identify the characteristics of centralized processing and distributed processing. For example, questions may be of the form “Which of the following is an appropriate characteristic of a centralized processing system?” Know the advantages and disadvantages of each processing type.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
104
2. Computer Systems
As shown in the following table, distributed processing can be classified according to the distribution status of functions and loads. It is said that vertical load distribution does not exist in reality. Configuration
Function
Horizontal distribution
Vertical distribution
Function distribution
Load distribution
Horizontal function distribution
Vertical function distribution
Horizontal load distribution
Horizontal function distribution
A horizontal function distribution system is a system in which computers are classified according to type of application and type of data; examples include processing function distribution and database distribution. For instance, in financial institutions, host computers may be classified into those in an information system and those in an accounting system; this classification is based on the type of processing, so it is an example of processing function distribution. Database distribution means that computers are located in appropriate locations based on the contents of data. Horizontal load distribution
This is a system in which multiple computers perform processes jointly when an application is executed. When a process is requested, an idle computer responds. In this mode, if one computer fails, the process switches to another computer and is continued. Hence, this system is quite effective in time of failure. A tightly coupled multiprocessor system is an example of this type.
Vertical function distribution
This is a system where the processing function is shared among workstations belonging to individual users as well as computers shared by multiple users. Here, there is a vertical relationship in regard to the processing function. A client/server system is a typical example of a vertical function distribution system.96
96
(Hints & Tips) A client/server system appears as if it were horizontal distribution, but it is properly classified under vertical function distribution. Since one server performs processes of multiple clients, there is a vertical relationship in functions. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
105
2. Computer Systems
2.3.4 Classification by Processing Mode
Points
¾
¾
In batch processing, data is stored and processed all at once.
In real-time processing, data is processed at the moment they come into existence.
From the standpoint of processing modes, system configurations can be classified into two categories: batch processing and real-time processing. They can also be classified by whether or not they are connected to a network.
Processing mode
Batch processing97
Real-time processing
Operation mode
Center batch processing
Remote batch processing
Interactive mode processing
Online transaction processing
Real-time control
Connection method
Offline
Online
Batch Processing Systems (One-Time Processing
Systems)
The word “batch” means a “bundle.” Batch processing is any method in which data to be processed by a computer are stored for a certain period of time or are saved on original sheets or storage media and are later processed all at once. Such systems have the following characteristics:
•
•
•
The computer can be used efficiently because the processing is done all at once.
It is suitable for routine and repetitive processing (standard tasks).
The results are not immediately obtained because the processing is collectively done.
In center batch processing systems, the processing takes place at a central computer center; in remote batch processing,98 the batch processing is performed from a terminal at hand via a communication line.
Interactive Processing Systems
This is a method in which the processing is performed in the mode of a dialogue with the computer. We can proceed with the processing while communicating with the computer and can correct errors immediately once they are noticed; therefore, it is a processing mode suitable for program development, etc. An interactive processing system requires prompt responses.
Also, an interactive processing system requires an OS in which the multiprogramming function is
97
Batch processing and real-time processing: Batch processing is where data is stored and processed all at once.
Examples include the calculations of electricity payment, water payment, and gas payment. Payroll calculation is also an example of batch processing. Real-time processing is where data is processed immediately when they are generated.
Seat-reservation systems of airlines and trains are examples of this type.
98
Center batch process and remote batch process: Center batch processing does not use communication lines. Remote batch processing is where a terminal unit at hand is used to perform batch processing at a host computer at a remote location. It is sometimes called RJE, which stands for “Remote Job Entry” as the user enters a job at a remote site.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
106
2. Computer Systems
supported, and the processing is performed using TSS.99
Online Transaction Processing Systems
A transaction is a series of data or instructions, sometimes called a transaction file or transaction data. Online transaction processing refers to the processing of data updates on an online file at the computer center via a connected online terminal. Examples of online transaction processing include savings account systems in banks and train and airline seat-reservation systems.
Online transaction processing requires specialized (dedicated) terminals with a high level of usability aimed at enhancing the processing efficiency, such as bank ATMs and terminals for issuing reserved-seat tickets at reservation windows. In addition, since data is shared by many terminals, there is a risk that simultaneous access to the same data could cause problems such as deadlock100 or data destruction. Hence, attempts are made to enhance the data maintainability.
Real-Time Control
In general, real-time control refers to the method by which data is processed in real time once a processing request is made and the result is immediately reported to the requester. This concept also includes online transaction processing. However, in a narrow sense, this refers more specifically to the processing mode at places like manufacturing plants, where the system is interlinked with a sensor system tracking the physical motions of objects to be controlled, processing corresponding to the external signals are immediately executed, and the results are immediately sent back to units on production lines (e.g. robots) as control signals. Production control systems at steel plants and automobile factories are examples of this type. Another example is a 24-hour monitoring process of electrical system managed by computers; if there is something wrong with the system, the system reports it to the maintenance company in real time, or the unit that has detected it gets cut off in real time.101
99
TSS (Time Sharing System): The system authorizes multiple programs to be executed in a specific order for an extremely short duration at a time (several milliseconds at a time). This procedure is repeated so that the execution of each program can get completed within a certain period of time. From the user's viewpoint, there is no mutual interference, so each user feels as if he or she were the only one exclusively using the computer system.
100
Deadlock: It is a situation where the system gets stuck because multiple tasks (programs) try to access the same resource (file, database, etc.) and go into the waiting mode.
101
(FAQ) Many exam questions involve characteristics of remote batch processing and interactive processing. Both are online, but note that remote batch is a type of batch processing while interactive processing is a type of real-time processing. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
107
2. Computer Systems
Quiz
Q1
Explain the roles of clients and servers in a client/server system.
Q2
What is the type of system configuration shown below?
Switching unit
Switching unit
(Primary system)
(Secondary system)
Q3
As shown below, distributed processing is classified by the distribution status of functions and loads. To which category does a client server system belong?
Configuration
Function
Horizontal distribution
Vertical distribution
Function distribution
Load distribution
Horizontal function distribution
Vertical function distribution
Horizontal load distribution
Q4
What is the difference between batch processing and real-time processing?
A1
Clients: They send service requests to servers.
Servers: They perform the requests sent by clients and return the processing results to them. A2
Duplex system: Distinguishable points are the existence of two CPUs and the switching units.
A3
Vertical function distribution
A3
Batch processing: Data is stored and processed all at once.
A4
Real-time processing: Data is processed immediately when they are generated.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
108
2. Computer Systems
2.4 Performance and Reliability of
Systems
Introduction
To evaluate computer systems, various methods are available. While good performance (high processing speed) is important, fault-tolerance (high reliability) is also significant.
2.4.1 Performance Indexes
¾
Points
¾
Response time, throughput, and turn-around time are used for performance evaluation.
Instruction mix and benchmark are used for performance indexes.
To evaluate the comprehensive performance of computer systems, including their software and hardware, we can use various criteria such as response time, throughput, and turn-around time.
Indexes to evaluate performance, especially the hardware, include instruction mix and benchmark.
Terms Related to Performance Evaluation of Computer
Systems
In evaluating the performance of computer systems, it is important to calculate the processing time of jobs and programs.102
Response time
This is the amount of time between the completion of input at an input unit and the beginning of output at an output unit. For example, when a processing request is made at the keyboard of the computer, this time refers to the amount of time it takes until the result is shown on the display unit or until the printing begins. This is mainly used to evaluate the performance of an online system. Throughput (Processing capability)
This refers to the amount/number of jobs that can be processed by the computer system within a certain unit of time, or the amount of time required to process a certain job. This processing time includes the exclusive CPU time and process-waiting time such as preparation for I/O operation and clean-up time.
102
(Hints & Tips) There are instruction mixes and benchmarks for evaluating the performance of computers, and instruction mixes are for hardware evaluation. However, even if hardware is very fast, the entire system performance becomes poor if the performance of the OS is poor. Hence, the performance of hardware is often used only for reference.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
109
2. Computer Systems
Turn-around time (TAT)
Technically, this refers to the amount of time it takes information to make the rounds of the system.
In batch processing, this is the duration between submission of a program at the window and the time when the results are obtained. In business operations, this is the duration from the time when a client places an order to the time when the ordered product is shipped and reaches the client.
Instruction Mix103
An instruction mix is used to compare the performance of hardware in computer systems. Even if the hardware is fast, if the performance of the OS is poor, the performance of the entire system becomes inferior. An instruction mix is to use an average program and calculate the average instruction execution time per instruction and MIPS value,104 based on the execution frequency of each instruction.
Under these conditions, let us do some specific calculations of the MIPS value.
Instruction group
A
B
C
Execution speed (microsecond)
0.1
0.2
0.5
Frequency of appearance
40%
30%
30%
First, let us calculate the average instruction execution time. The execution speed of each instruction is expressed in microseconds (10-6). The average instruction execution time is the sum (over all instructions) of the products of the execution time of instructions and their respective frequencies. Average instruction execution time = 0.1 * 10-6 * 0.4 + 0.2 * 10-6 * 0.3+0.5 * 10-6 * 0.3
= (0.04 + 0.06 + 0.15) * 10-6
= 0.25 * 10-6 (seconds/instruction)
The average number of instructions executed per second is the inverse of the average instruction execution time, so it is obtained as follows:
Average number of instructions executed per second = 1 / (0.25 * 10-6)
= 4 * 106 (instructions/second)
= 4 (MIPS).105 106
FLOPS107 is used as an index to evaluate the performance of floating-point operations.
103
(Note) An instruction mix for scientific calculations is called “Gibson mix,” and one for business calculations is called
“commercial mix.”
104
MIPS (Million Instructions Per Second): This is the performance index expressing the number of machine instructions, in millions (106), that can be executed per second. This is just for the performance of hardware, so again it is used only for reference.
105
(FAQ) Exams do have questions where you are asked to calculate MIPS values given an instruction mix or to calculate the average clock count per instruction. You would want to be familiar with these calculation questions.
106
Clock: This refers to the frequency of a clock signal generated by a circuit called a clock generator. Since instructions inside the CPU are synchronized to this clock signal as they are executed, the higher the clock frequency is, the more instructions can be executed in a given period of time. For example, if the clock frequency is 200MHz, there are 200 * 106 clock signals per second. In general, one instruction takes several clocks.
107
FLOPS: It stands for floating-point operations per second. This is an index expressing the number of floating-point operation instructions executed per second. If it is expressed in millions (106), it is called MFLOPS.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
110
2. Computer Systems
Benchmark
A benchmark is used to compare and evaluate the comprehensive performance of computers, including the hardware and the OS, by measuring the standard program execution time.108
2.4.2 Reliability
¾
¾
Points
Reliability indexes include RAS, RASIS, and the bathtub curve.
Points to be aware of in reliability design are fail-safe and fail-soft.
The level of reliability required for information systems varies depending on the purpose for which the systems are used. Sometimes the economical factor must be sacrificed to achieve a high level of reliability. In some other situations, not only the subject of reliability is focused on the operation of the system but also the information handled by the system needs to be reliable as well.
Reliability Indexes
Reliability is the degree to which system operation is stable. The ideal case is that the system does not fail, but there is no system that does not ever fail.
RAS/RASIS
Both of the terms RAS and RASIS are acronyms of elements that help computer systems to operate in a stable manner. RAS stands for the first three elements of RASIS:
R
A
S
I
S
R: Reliability (Few faults)
A: Availability (High level of availability109)
S: Serviceability (Easy detection of fault parts)
I: Integrity (Consistency of data)
S: Security (Prevention of invalid access)
108
(Note) An example of a technical calculation benchmark is SPECmark, and an example of a transaction benchmark is
TPC. TPC-C is the frequently used benchmark under TPC which directly responds to actual business applications.
109
Availability: Availability refers to the probability that the system is maintaining its functions (operating) at any given time or the percentage of the duration when the functions are maintained during a certain period of time.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
111
2. Computer Systems
Bathtub curve
Failure rate →
The bathtub curve is used to illustrate the concept of hardware lifecycle. Hardware may fail during the initial period of its operation due to defective parts, etc., but the probability at which these failures occur decreases gradually as repairs and replacements are made. After that, because of wear and tear of various parts, the probability of failures increases, and eventually its life is determined to be over. This curve is shown below.110
Early failure period111
Stable failure period112
Wear-out failure period113
Time →
Reliability Design Points
A highly reliable system that can continue to operate even when some part of the system fails is called a fault-tolerant system. Common technologies for configurations of highly reliable systems include fail-soft, the function enabling the system to continue its operation, perhaps with lower performance or fewer functions, when a failure occurs, and fail-safe, the function enabling the system to operate safely by avoiding risky conditions when a failure occurs.
Fail-soft
This refers to the function in which, when a failure occurs, the failed part gets cut off and the system continues to operate, perhaps with a lower performance level (fall back114). In a duplex system, normally the two systems are independently processing data, but if one system should fail, the configuration would switch the processing to the other system and would carry on the processing. In addition, when a failure occurs in multiprocessors, the system continues its services by cutting off the failed processor. This too is a system configuration with fail-soft in mind.
Fail-safe
This refers to the function in which, when a failure occurs, the system locks its functions in a safe mode established in advance to control the extent of the impact of the failure.115 This is just like the measure where all railroad lights turn red when an accident has occurred. In system configurations where two systems compare the processing results of each other, such as in a dual system, when the compared results are different, the system in which a failure is determined to have occurred is cut off while the operation continues on.
110
(Note) The bathtub curve is so named because the graph showing the relationship between the failure probability and time resembles the shape of a bathtub.
111
Early failure period: It is a period of failures at the beginning of unit use. These failures become less frequent as time passes. 112
Stable failure period: The unit is stable during this period, with less frequent failures.
113
Wear-out failure period: A certain period of time has passed, and failures become more frequent during this period.
114
Fall-back: In a fail-soft computer system, processing continues at a lower level of functionality; this is called a fall-back or a fall-back operation.
115
(Hints & Tips) Fail-soft and fail-safe are similar words, so do not confuse them.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
112
2. Computer Systems
Fool-proof
This term refers to a measure that prevents an unintentional use of a program from causing a failure, especially when indefinitely many users use the same program. If one individual is using a particular program, the way the program is written does not create a major problem, but when there are indefinitely many users, how the program gets used is hard to predict.116
2.4.3 Availability
¾
Points
¾
MTBF is the time when the system is operating properly, and
MTTR is the time when it is being repaired.
Availability is the ratio of the time when the system is operating properly. One of the indexes in RASIS is “A” for availability, which means the operation rate. The availability is calculated using MTBF and MTTR as follows:
Availability:
A = (MTBF) / (MTBF + MTTR)
MTBF and MTTR
Suppose that the operation status of a computer system is as shown below:
Time →
Operating
Being repaired T1
D1
Operating
Being repaired Operating
Being repaired T2
D2
T3
D3
…
Operating
Being repaired Tn
Dn
MTBF (Mean Time Between Failures)
This is the average length of time that the system continues to operate without a failure. The larger MTBF is, the more reliable the system is. Therefore, this is used as an index of reliability
(“R” in RASIS).117
MTBF = (T1 + T2 + T3 + … Tn) / n
(Here, “n” is the number of intervals the system was operating without failure.)
116
(FAQ) There are exam questions concerning what each of the letters RASIS means as an index for computer system reliability. At least know what RAS stands for.
117
(Note) Functions that improve MTBF include error detection, automatic 1-bit error correction, instruction re-try, etc.
These are functions that prevent the computer system from coming to a stop. Functions that improve MTTR include log output. By looking up logs, the cause of failure can sometimes be identified. Remote maintenance also helps detect a failure promptly, enhancing MTTR.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
113
2. Computer Systems
MTTR (Mean Time To Repair)
This is the average length of time required for repair when a failure occurs. The shorter the repair time is, the better the system is. Therefore, it is used as an index of serviceability (“S” in
RASIS).
MTTR = (D1 + D2 + D3 + … Dn) / n
(Here, “n” is the number of intervals the system was operating without failure.)
Calculation of Availability
To calculate the availability, the serial connection and parallel connection sections must be calculated differently. The basic ideas are described below.118
Availability in a serial connection system
The availability of an entire serial connection system as shown here is the product of the availabilities of each unit. Here, P1, P2, and P3 are the availabilities of the respective units shown in the figure.
Unit 1
P1
Unit 2
P2
Unit 3
P3
Availability of the entire system = P1 * P2 * P3
Availability in a parallel connection system
Suppose that we have, as shown below, a system in parallel connection where the system operates as long as at least one of Units 1, 2, and 3 is operating. Here, the availability is calculated using the fact that the probability of the entire sample space is 1.
Unit 1
P1
Unit 2
P2
Unit 3
P3
Availability of the entire system
= 1 – (Probability that all units fail simultaneously)
= 1 – (Prob. that Unit 1 fails) * (Prob. that Unit 2 fails) * (Prob. that Unit 3 fails)
= 1 – (1 – P1) * (1 – P2) * (1 – P3)
118
(FAQ) There is always a question involving a calculation of availability. Make sure you understand correctly how to calculate it.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
114
2. Computer Systems
Availability in a system where serial and parallel connections are combined
Suppose that there is a system which operates if Unit 1 is operating AND at least one of Units 2 and 3 is operating. In this case, we consider that Unit 1 and the parallel section (Units 2 and 3) are serially connected.
Unit
P2
Unit
P1
Unit
P3
Availability
= (Availability of Unit 1) * {1 – (Prob. that Units 2 and 3 fail simultaneously)}
= (Availability of Unit 1) * {1 – (Prob. that Unit 2 fails) * (Prob. that Unit 3 fails)}
= P1 * {1 – (1 – P2) * (1 – P3)}
Let us consider more complicated configurations.119 Even though the two systems below may appear similar, the availabilities are different. Here, the letter α in the figure indicates the availability. [Configuration 1] The two parallel sections (inside the dotted lines) are serially connected.
α
α
α
α
Availability = {1 - (1 - α)2} * {1 - (1 - α)2} = α2 (2 - α)2
[Configuration 2] The two serially connected units (inside the dotted lines) are connected in parallel. α
α
α
α
Availability = 1-(1 - α2) * (1 - α2) = α2 (2 - α2)
119
(Hints & Tips) Note that similar configurations have different availabilities.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
115
2. Computer Systems
Quiz
Q1
Explain the meanings of the following terms: “MIPS,” “response time,” “throughput,” and “turn-around time.”
Q2
Explain the meaning of RASIS.
Q3
Explain the meanings of MTBF and MTTR.
Q4
Express the availability using MTBF and MTTR.
Q5
Calculate the availability for the entire system configuration shown below. A and B are units, each of which has an availability of 0.97. The entire system is assumed to be in operation if at least one of the units is operating.
A
B
A1
MIPS:
It is an acronym standing for “million instructions per second.” This indicates the number of instructions that can be executed in one second, expressed in millions (106). It is one of the performance indexes of hardware. Response time:
It is the length of the time interval from data transmission to the return of processing results. It is one of the performance indexes of an online system. Throughput:
It is an evaluation measure of the job-processing capability of a computer system within a given amount of time. It is one of the performance evaluation indexes of a computer system.
Turn-around time: It is the length of the time interval from a job request to the complete output. Mainly a concept used in batch processing, this is an index of system evaluation including its operation.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
116
2. Computer Systems
A2 RASIS is an acronym of elements that help computer systems operate in a stable manner.
R
A
S
I
S
R: Reliability (Few faults)
A: Availability (High level of availability)
S: Serviceability (Easy detection of fault parts)
I: Integrity (Consistency of data)
S: Security (Prevention of invalid access)
A3
MTBF: (Mean Time Between Failures) This is the average length of time that the system continues to operate without a failure.
MTTR: (Mean Time To Repair) This is the average length of time required for repair when a failure occurs.
A4
Availability:
A5
Note that the connection is parallel.
Availability = 1 – (1 – 0.97) * (1 – 0.97)
= 1 – 0.03 * 0.03
= 1 – 0.0009
= 0.9991.
A = (MTBF) / (MTBF + MTTR)
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
117
2. Computer Systems
2.5
System Applications
Introduction
Various systems have been developed using networks and databases. Close to our daily life are the Internet and database services (generally called commercial databases). Examples of applications of multimedia systems include 3D graphics.
2.5.1 Network Applications
¾
Points
¾
Uses of the infrastructure include the Web, the Internet, intranets, and extranets.
Application systems include Internet shopping, groupware, and debit cards.
Today, our information society has networks spanning all over the world like a gigantic web.
Systems using networks themselves and application systems with add-on values are available.
This is a rather new area, so there are not many exam questions on this topic, and they are relatively easy. Most of the questions simply require knowledge of the terms, so be sure to memorize them to improve your exam scores.
Uses of Infrastructure
“Infrastructure” means “foundation” or “basis.” In computer systems, this word refers to the foundation of software and hardware to form the systems. For instance, in network construction, various components such as communication lines, communication units, and the charge system of the communication lines are parts of what is known as the communications infrastructure.
Web (WWW: World Wide Web)
This is the information search system in the hypertext format, developed by researchers at
CERN.120 Since information distributed all over the world is mutually linked by this network using hypertext,121 a name meaning “global spider web” was given to it.
WWW is a mechanism on the server that records information in the form of an Internet homepage.
Software that accesses WWW and displays it on a screen is called a Web browser or simply a
“browser.”
The Internet
120
CERN (Conseil Europeen pour la Recherche Nucleaire) (European Council for Nuclear Research): It is a quantum physics research institute jointly funded and operated by 12 European counties, but generally it is known as the institute which developed WWW on the Internet. Its name has now been changed to Laboratoire Europeen pour la Physique des
Particules, but the abbreviation remains the same.
121
Hypertext: It is a structure in which pointers are placed within texts so that links can be made to jump from those pointers to other texts and pictures. To create a document in the hypertext format, one uses HTML (HyperText Markup
Language). To identify a WWW server address, we can use URL (Uniform Resource Locator).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
118
2. Computer Systems
It is a collection of networks all over the world connected together by TCP/IP. There is no government organization or designated organization managing it in an integrated manner. Instead, the technical support and resource management are done by volunteer organizations.
ARPANET,122 created by the United States Defense Department, set the foundations for the
Internet.
Intranets
An intranet is a company-wide network applying the technology of the Internet. Normally, a firewall123 is set up between the Internet and a company-wide network in order to prevent leakage of confidential information of the company.
Firewall
Internet
Company-wide network Extranets
An extranet is an intranet extended over numerous companies. In general, an extranet is built by connecting intranets to the Internet.
Intranet at
Company A
Intranet at
Company B
The Internet
Mobile communication
Mobile communication124 is an environment in which the network can be accessed from any location. Today, communication is the mainstream, so we can send and receive e-mails on the
Internet and obtain a variety of information all with one telephone.
Satellite communication
Satellite communication is a wireless communications system using a communications satellite.
A broadcasting station transmits (uplink) a huge amount of information to a stationary communications satellite located 36,000 km above the equator (in a stationary orbit), and the information is distributed all at once (downlink) to various receiving stations on the earth. A large amount of information can be transmitted to many points.
CATV
New types of services using CATV (Community Antenna TeleVision, or Cable TV) are being considered and discussed commercially, such as Internet connection, telephone services, experiments involving PHS (Personal Handy-phone System), and VOD (Video On Demand).
CATV is expected to be a major part of the infrastructure in the multimedia era.125
122
ARPANET (Advanced Research Project Agency Network): This is a nationwide computer network developed under the sponsorship of the Advanced Research Project Agency (ARPA) of the United States Department of Defense. It is the predecessor of the Internet.
123
Firewall: It is the mechanism which is located between the Internet and a company-wide intranet to manage data communication and to protect the internal network from external attacks and invalid access. The word could also refer to this functional role.
124
Mobile/mobile computing: “Mobile" refers to any information device that can be carried around, including cell phones, PHSs (personal handy-phone system), and notebook PCs. “"Mobile computing” refers to the mode of using any of these information devices to have access to the company network from the outside.
125
(Note) CATV began as a reparation facility in remote areas and a community facility in rural regions. Today, urban
CATV, which can provide broadcast services on many channels, is getting attention as a new-generation component of the
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
119
2. Computer Systems
Application Systems
A network application system is a social system using a network. Specifically it includes the following: Internet shopping
This is a system in which the user can shop at a virtual store set up on web pages. To make a payment, the shopper can use his or her credit card or go to a nearby convenience store to pay.
Groupware
This is software for communication within an organization or for information sharing. It has functions such as electronic mail, schedule sharing, document sharing, and workflow.
Debit cards
This is a service whereby a cash card issued by a bank can be used to make payments. The money for the payment is directly withdrawn from the bank account in real time.126
infrastructure. Coaxial cables are used for distribution so that high-quality images can be received.
126
Debit cards have been traditionally called bank POS; cash cards are used instead of cash.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
120
2. Computer Systems
2.5.2 Database Applications
Points
¾
¾
An example of a database application is a data warehouse.
Applications in business include corporate accounting, inventory management, document management, and sales support.
One type of database application system is a data warehouse. Application systems in which databases are applied in business include corporate accounting systems, 127 inventory management systems,128 document management systems,129 and sales support systems.130
Data Warehouse
A data warehouse is a company-wide database to support decision-making. The idea is to have a large amount of data stored, organized, and used to help make business decisions. Sometimes it is called an informational database.
Data Mining
This refers to a technology or method of drawing out tendencies, trends, correlations, and patterns necessary for management and marketing, through dialogues with a large amount of raw data.
Whereas a data warehouse normally analyzes various data based on some hypothesis, data mining discovers trends and patterns in order to establish the hypothesis.
Data Mart
A data mart is a database which stores data obtained from a data warehouse. The data stored in a data mart, is selected and summarized according to the purposes of a specific user group.
Whereas a data warehouse contains information for the entire company, a data mart has a relatively small amount of data tailored for the target users.
127
Corporate accounting system: It is a system in which the accounting procedures of a corporation are computerized in an attempt to make the accounting tasks more efficient and quicker and to obtain timely understanding of the business and managerial records.
128
Inventory management system: It is a system to keep the production (purchase) and demand in balance, managing the inventory such as products and raw materials kept by the company at an optimum amount. In a retail store such as a supermarket, the sales information entered at POS terminals is collected and analyzed so that the demands can be predicted and more products are automatically ordered, taking into account safe inventory volumes and optimum amounts to purchase. 129
Document management system: It is a system in which a corporation manages various types of documents and sources; document search is possible from a variety of fields such as the storage location or contents of the document. It is an attempt to make document management and document preparation more efficient, e.g. to avoid duplicate preparation of the same document.
130
Sales support system: It is a system that supports making sales plans and business plans, based on accumulated sales information. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
121
2. Computer Systems
OLAP
OLAP (OnLine Analytical Processing) is the concept of analytical application in which the end user discovers problems and solutions by directly searching and organizing a database; the goal is to achieve quick data access and to provide a function for easy analysis.131
OLTP
OLTP (OnLine Transaction Process) is the processing mode in which messages are sent to the host computer from multiple terminals connected online to the host computer, which, according to the message received, in turn performs the process including access to a series of databases and returns the process results immediately to the terminals.
Databases used by OLTP are called business databases or, sometimes, operational databases.
These are terms in contrast with informational databases.
Below, we compare informational databases with operational databases:
Item for comparison
Target task
Data addition
Data updating
Processing mode
Main users
Business type
Period of data retention
Informational databases
Supporting decision-making
Yes
Generally no
OLAP
Management staff
Non-standard tasks132
Long term
Operational databases
General business tasks
Yes
Generally yes
OLTP
General workers
Standard tasks
Short term
Application Systems
Various systems that use databases are developed. Today, it is not an overstatement to say that most of the systems in operation use databases. Some of the great advantages for using databases are as follows:
• Data can be easily accumulated.
• Data can be managed in an integrated manner.
• Data can be easily processed.
• Data can be easily searched.
131
(FAQ) There have been exam questions concerning data warehouses. Know accurately the meanings of data warehouses, OLAP, and OLTP.
132
Standard/non-standard tasks: Standard tasks are those for which processing procedures are fixed, such as daily business procedures and daily input of sales data. Non-standard tasks are those for which processing procedures vary case by case. Creation of analysis documents, for instance, requires different processes depending on the purpose of use, so it is considered non-standard.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
122
2. Computer Systems
2.5.3 Multimedia Systems
¾
Points
¾
Examples of multimedia usage include AI, 3DCG, and pattern recognition. Multimedia application systems include Internet broadcasting and
VOD (Video On Demand).
Multimedia refers to handling not just characters and text but also mixtures of still images, moving pictures, audio, and other communication media. The term also refers to devices and software used in multimedia communication.
Artificial Intelligence (AI)
Artificial intelligence refers to a system that performs inference processes such as an expert system 133 and a machine translation system. A typical computer does no more than manipulating numerical values and performing logical operations. AI, on the other hand, performs inference processes centered on manipulating character strings. It is said that Lisp and
Prolog are languages suitable for developing software using AI technologies.
3-Dimensional Graphics (3D Computer Graphics, or
3DCG)
3D computer graphics134 involves creating virtual 3D space inside the computer, placing solid models in this space, and moving them around. It is embedded in movies, games, and animations. VRML135 is used to develop 3DCG.
The following table compares artificial reality and virtual reality.
Type
Artificial reality (AR)
Virtual reality (VR)
Explanations
It is the technology of creating a virtual world inside the computer, with a sense of reality.
Special equipment is not necessary to experience the artificial reality.
It is the technology of creating a fictitious world and having people experience and feel that world as though it were real.
3D vision using dedicated display units and special input equipment are used. Examples: pre-experience of virtual surgery, flight simulator, etc.
133
Expert system: It is a system created with the knowledge base of various specialists (experts) in a variety of fields; given certain conditions, the system applies the knowledge based on certain rules so that problems can be solved as if they were solved by the experts.
134
Computer graphics (CG): It is the technology of creating images via computers, or images made by such technology.
There are methods where the computer processes already existing images, and there are other methods where the computer creates images themselves. The latter method is called CGI (computer-generated images).
135
VRML (virtual reality modeling language): language specifications to describe 3DCG used on the Internet.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
123
2. Computer Systems
Multimedia Application Systems
There are many systems that apply multimedia. Now that WWW has brought in a graphical environment through the Internet, a variety of multimedia contents are available for use.
Internet broadcasting
This is broadcasting using multimedia on the Internet. With the use of streaming distribution technology, 136 programs are broadcasted in real-time on the Internet. Compared with conventional broadcasting business, equipment costs much less, and global information transmission is possible. In addition, on-demand service137 can also be provided, so we can tune in whenever we wish to watch a particular program.
Internet broadcasting comes in various formats. On-demand broadcasting stores the contents on a server and distributes them per request from a user. Live broadcasting (Internet live) distributes live programs, such as concerts, simultaneously to multiple users.
Non-linear editing
This is a method of video editing where images are digitized and video is produced in free order using a computer. It is easy to correct images, switch the order in which they appear, and create a different version. Incidentally, the conventional method in which video images are dubbed in the order of their completion is called linear editing.
Video on demand
This refers to the service of instantly sending a video program requested by the viewer via, for example, bidirectional CATV. The service provider stores many video programs on its video server and distributes the one requested by the viewer.
A video server can respond to simultaneous access by many viewers, and programs are requested to be sent from the beginning. Hence, it is necessary to construct an image database and connect it to individual mobile terminals and household receivers via broadband communication lines such as cable, wireless, etc.
136
Streaming: It is the technology of reading data and playing the data back immediately. It enables Internet broadcasting and playback of contents without waiting time. For streaming distribution, the line speed must exceed the amount of data; however, Internet lines are generally slow, so normally the data is compressed to enable real-time transmission.
Conventionally, playback used to be time-consuming since the data had to be downloaded first and then played back. With streaming, however, playback is done while data is being received.
137
On-demand: It is a function to provide what is requested whenever requested.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
124
2. Computer Systems
Quiz
Q1
What is an intranet?
Q2
What is a data warehouse?
A1
It is a network within a company applying the technologies of the Internet. Normally, a firewall is set up between the Internet and a company-wide network in order to prevent leakage of confidential information of the company.
A2
It is a company-wide database to support decision-making. It is the concept of storing a large amount of data in an organized manner to be used to help make business decisions. It is sometimes called an informational database.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
125
2. Computer Systems
Question 1
Q1.
There is a system which manages the file area in units of blocks. Each block contains eight sectors, and one sector is 500 bytes. How many sectors in total would be assigned to store two files, one consisting of 2,000 bytes and the other of 9,000 bytes?
Here, the sectors occupied by management information, such as directories, can be ignored. a) 22
b) 26
c) 28
d) 32
Answer 1
Correct Answer:
d
Files are saved in units of 8 sectors. Eight sectors, as shown below, are 4,000 bytes.
8 * 500 = 4,000 (bytes)
Hence, if one block is less than 4,000 bytes, all 4,000 bytes are used.
Next, we find the number of sectors necessary for each of the 2,000-byte file and 9,000-byte files. Capacity required for the 2,000-byte file
= 2,000 / 4,000=0.5 (block)
Æ 1 block (= 8 sectors) required
Capacity required for the 9,000-byte file
= 9, 000 / 4,000=2.25 (blocks)
Æ 3 blocks (= 24 sectors) required
Hence, to save files of 2,000 bytes and 9,000 bytes, the total number of sectors allocated to the two files is 32 as shown below.
Number of sectors required = 8 + 24 = 32 (sectors)
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
126
2. Computer Systems
Question 2
Q2.
Which of the following is the appropriate term for the process of breaking down data and storing it on multiple hard disks, as shown in the figure below? Here, b0 to b15 represent the sequence in which data is stored on the data disk in units of bits, and p0 to p3 represent the parity used to identify disk failure.
Control unit b0 b1
b2
b3
p0(b0 to b3)
b4
b5
b6
b7
p1(b4 to b7)
b8
b9
b10
b11
p2(b8 to b11)
b12
b13
b14
b15
p3(b12 to b15)
Data disk 1
Data disk 2
Data disk 3
Data disk 4
Parity disk
a) Striping
c) Blocking
b) Disk caching
d) Mirroring
Answer 2
Correct Answer:
a
Striping is to distribute one block of data onto two or more disks and write simultaneously. By striping, each block can be read and written in parallel, so the input/output speed increases.
Striping is defined as a technology for RAID. As shown in the figure above, this configuration contains a disk dedicated to the parity; such a configuration is called RAID2, RAID3, or
RAID4.
b) Disk cache is placed between a hard disk and the main memory; it is a buffer (buffer memory) to improve the apparent speed of the hard disk.
c) Blocking is to handle each logical set of multiple records as one physical record (block).
d) Mirroring is to prepare multiple disks and write the same data onto separate disks simultaneously, i.e., a multi-disk configuration. If one of the disks fails, the operation continues with the remaining disks only. This configuration is called RAID1.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
127
2. Computer Systems
Question 3
Q3.
Which of the following is arranged in the order of the effective memory access speeds from fastest to slowest?
With cache or w/o cache?
Cache memory
Access time
(ns)
-
Main memory
Hit rate
(%)
-
Access time (ns)
A
w/o cache
B
w/o cache
-
-
30
C
with cache
20
60
70
D
with cache
10
90
80
a) A, B, C, D
c) C, D, A, B
15
b) A, D, B, C
d) D, C, A, B
Answer 3
Correct Answer:
b
Cache memory (buffer memory) is memory which is placed between the CPU and the main memory to adjust speed differences between the two. The effective access speed can be increased by adding high-speed buffer memory, and by reading and writing on this cache memory as much as possible.
Let tc be the access speed of the cache memory, tm be the access speed of the main memory, and h be the hit ratio. The effective memory access speed is then calculated as follows:
Effective memory access speed = tc * h + tm * (1 – h)
The hit ratio is the probability that the data to be read is in the cache memory. The higher the hit ratio is, the faster the effective memory access speed becomes.
For A through D, we need to calculate the effective memory access time. For A and B, there is no cache memory, so the access time of the main memory is the effective memory access time.
Here, we can consider h = 0. The numbers below indicate the order of each, from fastest to slowest (from the shortest effective memory access time to the longest).
A: 15 (ns)
(1)
B: 30 (ns)
(3)
C: 0.6 * 20 + (1 – 0.6) * 70 = 40 (ns) (4)
D: 0.9 * 10 + (1 – 0.9) * 80 = 17 (ns) (2)
Hence, if we arrange Memory A through D from fastest to slowest in terms of the effective memory access time, the order is “A, D, B, C.”
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
128
2. Computer Systems
Question 4
Q4.
When a certain file was copied from one directory to another on a hard disk in a PC, file fragmentation occurred. Which of the following is an appropriate description concerning this situation?
a) The fragmentation can be eliminated by performing a physical dump of the entire disk and then restoring the disk.
b) Access time will be longer even for some files other than the file in which the fragmentation occurred.
c) If the file in which the fragmentation occurred is copied again, the fragmentation in the copy destination may get worse, but it will never be eliminated.
d) Even if fragmentation has occurred, the size of the file is still the same as that of the original one.
Answer 4
Correct Answer:
d
Fragmentation means that this file saved on the hard disk could not secure one continuous area and thus is saved across multiple blocks. When fragmentation occurs, various parts of the hard disk must be accessed, reducing the processing efficiency. However, the file size does not change, as the file is simply saved in a divided manner.
a) If the disk is physically copied, the situation does not change. It must be copied logically.
b) One file was copied on the hard disk on which files had already been saved, so no other file except the one that was copied was affected. Physically nothing was changed.
c) The fragmentation will be solved if the copy destination has a continuous empty area whose size is larger than the file size.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
129
2. Computer Systems
Question 5
Q5.
The table shown below gives processing times for a CPU and I/O devices to execute 5 stand-alone tasks. Which task can be executed simultaneously with the “High” priority task so that the CPU idle time from task execution start to end can be zero? Here, each task uses a different I/O device and is performed concurrently. The overhead of the OS can be ignored.
Unit: ms
Priority
Stand-alone task processing time
High
CPU (3) →I/O (3) → CPU (3) → I/O (3) → CPU (2)
a)
Low
CPU (2) →I/O (5) → CPU (2) → I/O (2) → CPU (3)
b)
Low
CPU (3) →I/O (2) → CPU (2) → I/O (3) → CPU (2)
c)
Low
CPU (3) →I/O (2) → CPU (3) → I/O (1) → CPU (4)
d)
Low
CPU (3) →I/O (4) → CPU (2) → I/O (5) → CPU (2)
Answer 5
Correct Answer:
c
We check the operation of each “low-priority” task to see which one uses the CPU while the
“high-priority” task is not using the CPU. The input/output units are different, so there is no waiting for the input/output (I/O) units.
The “high-priority” task uses I/O units twice, each for 3 ms. If the use of the CPU by the
“low-priority” task takes exactly 3 ms, the CPU has no idle time. With this said, let us now consider each task listed in the answer group.
For instance, consider Task (a). During the 3-ms period when the “high-priority” task uses an
I/O unit for the first time, the “low-priority” task uses the CPU for 2 ms. Hence, 1 ms of CPU idle time will result.
Similarly, any “low-priority” tasks like (b) and (d), with 2 ms of CPU use, will cause CPU idle time. In contrast, Task (c) has 4 ms of CPU use, but this comes in the end, so no idle time will occur if this is after the completion of the “high-priority” task.
Hence, the “low-priority” task that can completely eliminate the CPU idle time until the execution of both tasks is completed is “CPU (3), I/O (2), CPU (3), I/O (1), CPU (4).”
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
130
2. Computer Systems
Question 6
Q6.
The state transition diagram below shows a task (process) state transition on a multitasking computer. When does the task state change from the running state to the ready state?
Running state
Ready state
a)
b)
c)
d)
Waiting state
A task with higher priority compared to its own has moved to the ready state.
A task has been generated by the job scheduler.
An I/O operation has completed.
An I/O operation has been requested.
Answer 6
Correct Answer:
a
When a task (process) is generated, it proceeds to the ready state and is entered into the queue.
After that, it goes into the running state depending on the task priority and goes into the waiting state when an I/O operation occurs. Such changes in the task state are called the state transitions of the task.
(5)
(1)
(3)
task generated Running state B
A
Task eliminated C
E
Ready state F
D
Waiting state (4)
(2)
We now explain the state transitions of tasks. The numbers in parentheses ( ) are for state explanations, and A through D are for transition explanations.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
131
2. Computer Systems
No.
(1)
(2)
(3)
(4)
(5)
State
Task generated
Ready
Running
Waiting
Task eliminated
Transition
A
B
C
D
E
F
Explanation
A task generated is registered into the task queue.
Waiting to be authorized to use the CPU
Authorized to use the CPU and is currently being executed
Waiting for completion of an event such as an I/O operation
A task is completed and no longer necessary; it is removed from the task queue.
Explanation
A task is generated; moved to the ready state
By priority, moved to the running state
Moved to the waiting state (waiting for I/O, etc.) due to an event
Moved to the ready state after completion of an event
Moved to the ready state due to another task with high priority
The task in the running state is completed.
As you can see from these tables, transition from the running state to the ready state is transition E. This is when a task whose priority is higher than the task being executed goes into the ready state.
b) This describes transition A in the state transition figure for the tasks.
c) An I/O operation is an operation of input or output. When this I/O operation is completed, an
I/O interruption occurs and the task moves into the ready state. This describes transition D in the state transition figure for the tasks.
d) An I/O operation is requested when the task issues an instruction for input or output. If this happens, a supervisor call occurs, and the task moves from the running state to the waiting state to wait for the completion of the I/O. This is transition C in the state transition figure for the tasks.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
132
2. Computer Systems
Question 7
Q7.
Which of the following is an appropriate statement concerning a client/server system?
a) The client and the server must use the same kind of OS.
b) The server sends data processing requests and the client processes those requests.
c) A server can support a client function that enables it to request processing of another server if necessary.
d) The server functions must be allocated to different computers, such as a file server and print server.
Answer 7
Correct Answer:
c
A client server system is a system made up of processing units called clients and servers. A client performs data input/output and other processes through a server while a server controls all input and output that depend on the hardware according to its type. Normally, a client unit is a unit equipped with data processing functions such as a personal computer or a workstation, so it can perform applications on its own. In fact, it also performs processes only a client can perform, such as displaying text and drawing figures. A server, on the other hand, accesses databases and performs printing processes in response to requests by clients. Further, if a server is not able to perform a process, it can request another server to do that. Here, the first server becomes a client because it is requesting another to perform a process.
a) Different operating systems do not cause any problems as long as the protocol is established.
We can have a combination of servers with UNIX and clients with Windows.
b) It is a client that sends requests for processing. It is a server that performs the requested processes. d) In a small-scale system, a server and a client can even be the same. X Windows of UNIX is an example of this type. If the clients and servers are built on the same platform (OS) and can be connected via a network, there is no inhibitory effect.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
133
2. Computer Systems
Question 8
Q8.
When comparing a distributed processing system, which consists of multiple computer systems located in a wide area, with centralized processing systems that operate in a single center, which of the following is the most appropriate feature of centralized processing systems?
a) In the event of a disaster or a failure, recovery work can be conducted in a centralized fashion in the center, avoiding the risk of a long shutdown of the entire system. b) Since the system is collectively managed, it is easy to satisfy requests for additions or changes to system functions, and the accumulation of backlog seldom occurs.
c) Data consistency is easily maintained and managed through the centralized implementation of measures in the center.
d) Although the operation and management of hardware and software resources become complex, expansion taking advantage of new technologies is easy.
Answer 8
Correct Answer:
c
A centralized processing system is the idea of centralizing processes to a general-purpose large computer (host computer). Since the equipment is centralized at one location, the management is easy. However, if the host computer fails, it is possible that the system gets shut down for a long period of time.
a) Since the host computer performs all processing, the system gets shut down until the host computer is recovered. If this is a serious failure, it is possible that the downtime of the system becomes long.
b) In a centralized processing system, all requests must be answered by the host computer alone.
If the contents of the requested items vary significantly in level, the host computer cannot meet all those requests, causing backlog accumulation. Incidentally, backlog can also refer to a system waiting to be developed.
d) A centralized processing system processes all tasks at the host computer, so it is cumbersome to respond to each type of business task separately. Even if we want to extend the system, often there are tasks that cannot be suspended. With the introduction of a new technology, some tasks may not be able to be processed any longer.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
134
2. Computer Systems
Question 9
Q9.
For one job, which of the following formulas appropriately expresses the relationship between turnaround time, CPU time, I/O time, and process waiting time? Here, other types of overhead time are ignored.
a)
b)
c)
d)
Process waiting time = CPU time + Turnaround time + I/O time
Process waiting time = CPU time - Turnaround time + I/O time
Process waiting time = Turnaround time - CPU time - I/O time
Process waiting time = I/O time - CPU time - Turnaround time
Answer 9
Correct Answer:
c
Process waiting time is the time until the start of a CPU process or an I/O process of the job entered into the computer system. Turn-around time (TAT) is the time interval from submitting a job to receipt of the results. This concept is mainly used in batch processing. TAT includes both the CPU processing time and I/O time. Hence, process waiting time is obtained by subtracting CPU processing time and I/O processing time from TAT.
In the figure below, the shaded areas represent process waiting time.
Job entered
Job completed
CPU
processing
I/O processing CPU processing I/O processing Turn-around time
Hence, for one job, the formula expressing the relationship among turn-around time, CPU time,
I/O time, and process waiting time is as follows:
Process waiting time = Turn-around time – CPU time – I/O time.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
135
2. Computer Systems
Question 10
Q10. LAN facilities are installed as shown in the figure below. Using the server connected to LAN3, the client on LAN1 is performing a business application. Data transmission is normally performed via Router 1. If a failure occurs in Router 1, Routers 2 and 3 are used for transmission between LAN 1 and LAN 3. What is the availability of the
LAN equipment connecting LAN1 and LAN3? Here, the failure rate of each router is
0.1, no switch-over time is required in case of a failure, and failures in LAN facilities other than the routers are not taken into account.
Server
LAN 3
Router 3
LAN 2
Router 1
Router 2
LAN 1
Client
a) 0.729
b) 0.981
c) 0.990
d) 1.000
Answer 10
Correct Answer:
b
Note that “Router 1” and “the serial connection of Routers 2 and 3” are in parallel connection.
Since the failure rate of each router is 0.1, the availability is 0.9. Hence, the configuration is as shown below. The values in the boxes are availability of each unit.
Configuration 1
Router 2
0.9
Router 3
0.9
Client
Server
Router 1
0.9
Availability of Configuration 1 = 0.9 * 0.9 = 0.81
Availability of the entire system = 1 – (1 – 0.81) * (1 – 0.9) = 0.981.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
136
2. Computer Systems
Question 11
Q11. Which of the following is an appropriate statement concerning VR (Virtual Reality)?
a) Using technology such as CG, VR expresses the world created inside a computer as if it were the real world.
b) For the purpose of improving GUI, it does not display an image incrementally from the top, but first displays a rough mosaic-like image and gradually sharpens it.
c) VR tests whether or not hypothetical results can be obtained from computer simulations such as those of wind tunnel tests used for automobile or aircraft design. d) VR makes abilities such as human recognition and inference possible on a computer. Answer 11
Correct Answer:
a
Virtual reality (VR) is to artificially create a sense of reality by combining CG (Computer
Graphics) and sound effects. To appeal to the senses, we may also use dedicated display units such as a head-mounted display (HMD) for 3D vision and special input units such as data gloves. In addition, the response to the images is returned to the user, further enhancing the sense of reality.
b) This describes interlace.
c) This describes simulation.
d) This describes AI (Artificial Intelligence).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
137
FE Exam
Preparation Book
Preparation Book for Fundamental Information Technology Engineer Examination
Part1: Preparation for Morning Exam
Part2: Trial Exam Set
INFORMATION-TECHNOLOGY PROMOTION AGENCY, JAPAN
FE Exam Preparation Book
Vol. 1
Table of Contents
Part 1
PREPARATION FOR MORNING EXAM
Chapter 1
Computer Science Fundamentals
1.1 Basic Theory of Information
1.1.1 Radix Conversion
1.1.2 Numerical Representations
1.1.3 Non-Numerical Representations
1.1.4 Operations and Accuracy
Quiz
1.2 Information and Logic
1.2.1 Logical Operations
1.2.2 BNF
1.2.3 Reverse Polish Notation
Quiz
1.3 Data Structures
1.3.1 Arrays
1.3.2 Lists
1.3.3 Stacks
1.3.4 Queues (Waiting lists)
1.3.5 Trees
1.3.6 Hash
Quiz
1.4 Algorithms
1.4.1 Search Algorithms
1.4.2 Sorting Algorithms
1.4.3 String Search Algorithms
1.4.4 Graph Algorithms
Quiz
Questions and Answers
i
2
3
3
7
10
11
14
15
15
18
21
24
25
25
27
29
30
32
34
37
38
38
41
45
48
50
51
Chapter 2
Computer Systems
62
2.1 Hardware
2.1.1 Information Elements (Memory)
2.1.2 Processor Architecture
2.1.3 Memory Architecture
2.1.4 Magnetic Tape Units
2.1.5 Hard Disks
2.1.6 Terms Related to Performance/ RAID
2.1.7 Auxiliary Storage / Input and Output Units
2.1.8 Input and Output Interfaces
Quiz
2.2 Operating Systems
2.2.1 Configuration and Objectives of OS
2.2.2 Job Management
2.2.3 Task Management
2.2.4 Data Management and File Organization
2.2.5 Memory Management
Quiz
2.3 System Configuration Technology
2.3.1 Client Server Systems
2.3.2 System Configurations
2.3.3 Centralized Processing and Distributed Processing
2.3.4 Classification by Processing Mode
Quiz
2.4 Performance and Reliability of Systems
2.4.1 Performance Indexes
2.4.2 Reliability
2.4.3 Availability
Quiz
2.5 System Applications
2.5.1 Network Applications
2.5.2 Database Applications
2.5.3 Multimedia Systems
Quiz
Questions and Answers
ii
63
63
65
68
70
73
77
79
81
83
85
85
87
89
90
95
99
100
100
102
104
106
108
109
109
111
113
116
118
118
121
123
125
126
Chapter 3
System Development
138
3.1 Methods of System Development
3.1.1 Programming Languages
3.1.2 Program Structures and Subroutines
3.1.3 Language Processors
3.1.4 Development Environments and Software Packages
3.1.5 Development Methods
3.1.6 Requirement Analysis Methods
3.1.7 Software Quality Management
Quiz
3.2 Tasks of System Development Processes
3.2.1 External Design
3.2.2 Internal Design
3.2.3 Software Design Methods
3.2.4 Module Partitioning Criteria
3.2.5 Programming
3.2.6 Types and Procedures of Tests
3.2.7 Test Techniques
Quiz
Questions and Answers
139
139
141
143
144
147
149
151
154
155
155
157
159
162
163
165
167
170
172
Chapter 4
Network Technology
181
4.1 Protocols and Transmission Control
4.1.1 Network Architectures
4.1.2 Transmission Control
Quiz
4.2 Transmission Technology
4.2.1 Error Control
4.2.2 Synchronization Control
4.2.3 Multiplexing and Communications
4.2.4 Switching
Quiz
4.3 Networks
4.3.1 LANs
4.3.2 The Internet
4.3.3 Various Communication Units
4.3.4 Telecommunications Services
Quiz
Questions and Answers
iii
182
182
184
187
188
188
190
192
194
195
196
196
198
200
202
204
205
Chapter 5
Database Technology
212
5.1 Data Models
5.1.1 3-layer Schemata
5.1.2 Logical Data Models
5.1.3 E-R Model and E-R Diagrams
5.1.4 Normalization and Reference Constraints
5.1.5 Data Manipulation in Relational Database
Quiz
5.2 Database Languages
5.2.1 DDL and DML
5.2.2 SQL
Quiz
5.3 Control of Databases
5.3.1 Database Control Functions
5.3.2 Distributed Databases
Quiz
Questions and Answers
213
213
215
217
218
221
223
224
224
226
231
232
232
234
236
237
Chapter 6
Security and Standardization
244
6.1 Security
245
245
6.1.1 Security Protection
6.1.2 Computer Viruses
247
6.1.3 Computer Crime
249
Quiz
251
252
6.2 Standardization
6.2.1 Standardization Organizations and Standardization of Development and
Environment
252
6.2.2 Standardization of Data
254
6.2.3 Standardization of Data Exchange and Software
256
Quiz
258
259
Questions and Answers
iv
Chapter 7
Computerization and Management
7.1 Information Strategies
7.1.1 Management Control
7.1.2 Computerization Strategies
Quiz
7.2 Corporate Accounting
7.2.1 Financial Accounting
7.2.2 Management Accounting
Quiz
7.3 Management Engineering
7.3.1 IE
7.3.2 Schedule Control (OR)
7.3.3 Linear Programming
7.3.4 Inventory Control (OR)
7.3.5 Probability and Statistics
Quiz
7.4 Use of Information Systems
7.4.1 Engineering Systems
7.4.2 Business Systems
Quiz
Questions and Answers
Part 2
262
263
263
265
267
268
268
270
274
275
275
278
282
284
286
290
291
291
293
296
297
TRIAL EXAM SET
Trial Exam Set
309
Fundamental IT Engineer Examination(Morning) Trial
Answers and Comments
Fundamental IT Engineer Examination(Afternoon) Trial
Answers and Comments
v
310
339
381
419
2
Computer Systems
Chapter Objectives
A computer system is composed of hardware and software. There are many types of computer, but the principles of their operation are fundamentally the same.
We will learn the mechanism of computers (hardware) in
Section 1 and software (operating system) for efficient computer use in Section 2. We will further learn some configurations of computer systems for achieving improved reliability in Section 3 and ways to evaluate the performance of computers in Section 4. Finally, in Section
5, we will learn various systems that use computers.
2.1
2.2
2.3
2.4
2.5
Hardware
Operating System
System Configuration Technology
Performance and Reliability of Systems
Systems Application
[Terms and Concepts to Understand]
Central Processing Unit (CPU), cache memory, input/output interface, auxiliary memory, task management, job management, multiprogramming, virtual memory, dual system, duplex system, client/server system, availability, MTBF, MTTR, Internet
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
62
2. Computer Systems
2.1 Hardware
Introduction
Hardware is defined as devices with which a computer is configured. Computers consist of processing units, memory, input/output units, etc. In this section, we will explain these units from the standpoint of hardware.
2.1.1 Information Elements (Memory)
Points
¾
¾
Information elements include ROM and RAM.
SRAM and DRAM are typical types of RAM.
Semiconductor memory is memory made of integrated circuits (ICs) using semiconductors.
Semiconductor memory includes ROM, which is non-rewritable, and RAM, which is rewritable. ROM (Read-Only Memory)
ROM is semiconductor memory that is not erased when the power is turned off.1 On mask
ROM and PROM, data can be written only once; on EPROM, however, data can be written and re-written repeatedly using a special method. Types and characteristics of ROM are shown in the following table.
Type (Name)
Mask ROM
PROM (Programmable ROM)
EPROM (Erasable PROM)
EEPROM (Electrically EPROM)
Flash memory2
Characteristics, etc.
Data is written at the time of manufacturing. It cannot be re-written later.
PROM data is written by the user when it is first used. It cannot be re-written later.
EPROM data is written by the user electrically. All the data can be erased using ultraviolet rays.
All the data can be erased and re-written. Data is erased electrically. Erasure and re-writing can be done collectively or on a block basis. Data is erased electrically.
1
Volatility and non-volatility: It is the property that the contents of memory are lost when the power is turned off is called volatility. RAM is a type of volatile memory. On the other hand, the property that the contents of memory are not lost when the power is turned off is non-volatility. ROM is a type of non-volatile memory.
2
(Hints & Tips) Flash memory is classified as EEPROM.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
63
2. Computer Systems
RAM (Random Access Memory)
RAM is semiconductor memory that loses its memory contents when the power is turned off.
Unlike ROM, its contents can be changed, so it is used for main memory, graphics memory,3 and cache memory.
There are two typical types of RAM: SRAM and DRAM. The characteristics of SRAM and
DRAM can be summarized as shown in the following table.
Comparison item
Level of integration
Access speed
Price
Usage
Operation
Structure
SRAM
Low (small capacity)
Fast
Expensive
Cache memory4
Battery-operated devices
No refresh is required.
Flip-flop
Complicated structure
DRAM
High (large capacity)
Slow
Inexpensive
Main memory
Refresh is required.
Condensers and transistors
Simple structure
SRAM (Static RAM)
SRAM is composed of a flip-flop,5 so it does not require any refresh operations and is able to speed up information reading and writing. However, the cost is higher for the same capacity than DRAM, because the SRAM structure is more complicated than that of DRAM. For this reason, it is used mainly in areas where the speed, not the cost, is important, such as in cache memory. It is also used in battery-operated devices.
DRAM (Dynamic RAM)
DRAM consists of condensers and transistors, representing whether or not there is electrical charge in the condensers, by using 1 or 0. As time elapses, the electrical charge in the condensers gets discharged, resulting in memory loss; therefore, it needs to be re-written
(refreshed) at certain time intervals (every few milliseconds). Since the structure is rather simple, the manufacturing cost is low, and it is mainly used in the main memory of PCs.6
Types of DRAM equipped with high-speed data transfer functions include SDRAM, DDR
SDRAM, etc.
3
Graphics memory: It is memory used when images and characters are displayed on the display screen using a computer.
It is also referred to as video memory (VRAM).
4
Cache memory: It is high-speed memory placed between main memory and the CPU to speed up data reading from main memory to the CPU.
5
Flip-flop (also know as bistable circuit): It is an electrical circuit with two stable states, which maintains its state until an input that changes one of the states is entered.
6
(FAQ) Frequently there are questions that compare SRAM and DRAM. You should have good knowledge of the level of integration, usage, structure, etc.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
64
2. Computer Systems
2.1.2 Processor Architecture
Points
¾
¾
A computer consists of five major units (functions).
There are several addressing methods; direct addressing, indirect addressing, etc.
The term architecture refers to “structures or organizations.” The processor architecture refers to the configuration and operating principles of the computer.
Configuration of Computer
Below is a figure showing the basic configuration of a computer. This configuration is called the “big five units” or “big five functions,”7 because there are five major components.
Processing unit
Control unit
Control flow
Data flow
Operation unit
Input unit
Memory
Main memory Output unit
Auxiliary memory The control unit and the operation unit are together called the processing unit or the central processing unit (CPU).
Address Modification and Addressing Methods
A program is stored in the main memory and is retrieved, one instruction at a time, by the control unit to be executed. Address modification occurs in order to locate the location of data that is subject to processing.
Address modification is a function that obtains the value of the address actually accessed based on the address specified by the instruction. The method for address modification is called an addressing method. The address actually accessed as the result of the address modification is called the effective address. Addressing methods is described below in detail.
7
•
•
•
•
•
“The big five units”:
Control unit: It is the unit that controls the entire computer. It extracts and reads instructions of the program stored in the main memory and sends to various units the directions necessary to execute the instruction.
Operation unit: It is the unit that performs the arithmetic operations, logic operations, and other operations. It consists of adders, registers, complementers (units that convert values to their complements), etc.
Memory: It is a generic term of the unit that stores data, programs, etc. It can be classified into main memory and auxiliary memory.
Input unit: It is a generic term of the unit that enters programs and data into the computer.
Output unit: It is a generic term of the unit that outputs results of computer processing in characters and numbers that we can recognize.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
65
2. Computer Systems
Direct addressing method
In this method, the content stored in the address part of the instruction becomes the data subject to operation.
（Direct addressing method）
Instruction
part
Address part x Address
Main memory
∼
0
1
2 x
Indirect addressing method
In this method, the data stored in the address designated by the address part of the instruction are not the data subject to operation; rather, the data stored at the address designated by that content are the data subject to operation.
（Indirect addressing method）
Instruction
part
Address part x Address
Main memory
0
1
x
y
y
Indexed addressing method (index modification)
In this method, the effective address is the sum of the value of the address of the instruction and the value of the index register.8 For example, when processing an array, we can look up the content of another address simply by changing the content of an index register.9
（Indexed addressing method）
Instruction
part
i
Address part x Address
Index
Index register
0
1
0
1
x
y
8
Main memory y
x+y
Register: It is low-capacity, high-speed memory where data is temporarily stored. It is located in the CPU. There are various registers, including the following: general registers, for storing intermediate and final results of operations; status registers for indicating the CPU state after an instruction is executed; index registers for address calculations; and base registers. 9
(FAQ) There are questions on the concept of addressing methods. An example is “Which of the following is an appropriate description of the direct addressing method?” Be sure to have these methods organized in your mind: the direct addressing method, index addressing method, immediate value addressing method, etc.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
66
2. Computer Systems
Base addressing method
In this method, the effective address is the sum of the address designated by the address part of the instruction and the content of the base address register.10
（Base addressing method）
Instruction
part
B
Address part x Address
Index
Base register
0
1
B
0
1
x
b
Main memory y
b+y
Immediate value addressing method
This is the method where the address of the instruction stores the data subject to processing, not an address.11
RISC and CISC
Various types of technology have been used to speed up computer processing time; RISC and
CISC are examples of this technology. With the advancement of semiconductor technology, the integration density of integrated circuits has risen continually. At present, computers are made with integrated circuits, and RISC and CISC are the two approaches for developing computers.
Computers configured for high speed with simple instructions and simplified hardware are called RISCs. In contrast, those where complicated instructions are configured on one circuit are called CISCs.
RISC (Reduced Instruction Set Computer)
These computers have only a set of simple, frequently used instructions integrated onto a single
VLSI (very large scale integration) chip in order to achieve high performance through improved machine cycles (operation speed) and a reduction in instruction processing time. The emphasis is placed on keeping the length of each instruction to a fixed length and limiting the time required to execute each instruction to a fixed amount. By doing so, the technology of pipeline control has been easily implemented. However, the number of instructions to be executed becomes large unless efficient object programs are created, so it is essential that the compiler have an optimization function.12 Most computers called workstations are of this type.
CISC (Complex Instruction Set Computer)
These computers have complex instructions integrated onto a single VLSI ship in order to achieve high overall performance. Most general-purpose computers are CISCs.
10
(Note) The base addressing method can be used regardless of where in the main memory the program is stored, simply by changing the value of the base register. Such a structure is called a re-locatable structure.
11
(Hints & Tips) Note that in the immediate value addressing method, the address of the main memory is not designated.
12
Optimization: It is a function of a compiler to eliminate redundancy of a program in order to reduce the execution time of the object program and the size of the program. This is done in a variety of ways, such as calculating constants in advance, simplifying formulas, and eliminating double loops.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
67
2. Computer Systems
Pipeline Control
We have mentioned RISC and CISC as technologies to improve computer processing speed. To further improve the speed, the RISC system uses pipeline control. Pipeline control is a technology to reduce the instruction execution time of the CPU. This is an attempt to do the following: when execution steps of an instruction are divided into 5 or 6 steps, and if each step is completed within a certain fixed amount of time and the instruction steps stay independent of one another, then we can improve the overall processing speed by delaying the execution of each instruction 1 step behind the previous instruction. In reality, however, due to branching instructions, there are times when the next instruction address is not completely determined, and some steps are not completed within the fixed processing time. Everything is not always functioning in an ideal way, but pipeline control does process instructions concurrently, providing one way of speeding up the computer.13
Execution order
Time
Instruction 1
I1
Instruction 2
I2
I3
I4
I5
I1
I2
I3
I4
I5
I1
I2
I3
I4
I5
I1
I2
I3
I4
Instruction 3
Instruction 4
I1 : Fetch an instruction
I2 : Decode
I3 : Calculate an address
I4 : Fetch data
I5 : Execute
I5
2.1.3 Memory Architecture
¾
Points
¾
A hierarchical memory structure is introduced in order to achieve high-speed, large-capacity memory.
Cache and interleaving technologies are used for speeding up memory. There are many requirements for memory, but requirements for high speed and large capacity14 are of particular importance. In general, however, high-speed memory is expensive and has small capacity whereas low-speed memory is inexpensive and has large capacity. So, efforts are being made to combine high-speed but small-capacity memory and low-speed but large-capacity memory to develop high-speed and large-capacity memory.
13
(FAQ) Questions regarding pipeline control often appear on exams. Most of them are in the form of choosing an appropriate description of pipeline control, so you only have to know that pipeline control executes instructions concurrently. 14
(Note) Besides these, requirements for memory include reliability, ability of random access, non-volatility, re-writable function, portability, low cost, etc.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
68
2. Computer Systems
Memory Hierarchy
Memory hierarchy is a hierarchical representation of the relationship between the access speed and capacity of various types of memory.15
Expensive
price
speed
Fast
Inexpensive
Slow
Register
Cache memory16
Main memory
Disk cache
Auxiliary memory (hard disk, etc.)
Large-capacity memory (optical disk, etc.)
Memory capacity
Cache Memory (High-Speed Buffer Memory)
Cache memory is high-speed, small-capacity memory that is placed between the CPU or registers and the main memory. The main memory is slower than the CPU or registers, so the
CPU process can be made more efficient by storing frequently accessed data and programs of the main memory into the cache memory. Recently, a secondary cache has been installed for further improvements in speed.
The cache memory placed between a hard disk and the main memory is called the disk cache.
The figure below shows the relationship between the cache memory and the disk cache.
Central Processing Unit (CPU)
Cache memory
Primary cache
Secondary cache17
Main memory
Disk cache
Hard disk unit
15
(Note) If t is the average memory access time, tm is the access time of the main memory, tc is the access time of the cache memory, and h is the hit ratio, then the following equation holds: t = tch + tm(1 – h).
16
Hit ratio: It is the probability that the portion of a program necessary to execute that program is in the cache memory
17
Secondary cache: The primary cache is the cache memory which is built in the CPU; the secondary cache is the cache memory placed between the primary cache and the main memory.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
69
2. Computer Systems
Interleaving (Memory Interleaving)
The main memory is divided into multiple units called banks, and addresses are assigned across the banks. Often, the main memory is accessed over a sequential range of addresses at a time, so the speed can be enhanced by accessing a sequential range of addresses concurrently. For example, even when the operation of Bank 0 is not yet completed, Bank 1 can be accessed.
Below is an example of 4-way interleaving (with 4 banks).
Bank 0
0
4
Address
Address
Bank 1
1
5
Bank 2
2
6
Bank 3
3
7
…
…
…
…
Central Processing Unit (CPU)18
Data and programs are stored over a sequence of addresses (horizontally), but the memory is accessed in bank units (vertically). This allows concurrent access to a sequence of addresses.
2.1.4 Magnetic Tape Units
Points
¾
¾
Capacity calculations require blocking factors and record density.
Performance calculations do not include the stop time.
Magnetic tape is a medium that records data onto a tape that has been magnetically coated.
The unit price of this memory medium is cheap and has a large capacity, so it is used in cases such as backing up entire hard disks.
Capacity Calculation
The record format of a magnetic tape is shown below. As we can see in this figure, in order to record one block, we needs to include an IBG (Inter-block gap) which is an area to identify the block and contains a special code. A magnetic tape reader reads data in block units, so this area is crucial even in identifying the end of each block.
Record
length
…
IBG
Record 1
Record 2
Record 3
…
Record n
IBG
…
Block length
Tape length necessary to record one block
18
(FAQ) Interleaving is a way to speed up memory. Questions on the concept of interleaving have appeared often, so be sure you understand this.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
70
2. Computer Systems
Assuming that the specifications of a magnetic tape are given below, let us calculate precisely the number of records that can be stored on this single magnetic tape.19
[File specifications]
Record length
Blocking factor
[Magnetic tape specifications]
Record density
Inter-block gap (IBG)
Tape length
80 bytes
100
64 bytes/mm
15 mm
730 m
Calculation of block length
The blocking factor is 100. Æ 100 records can be stored in one block.
Record length Æ 80 bytes
So the number of bytes L1 for each block, excluding the IBG, is as follows:
L1 = 80 (bytes/record) * 100 (records/block) = 8,000 (bytes/block).
The record density is 64 bytes/mm. Æ 64 bytes can be stored on 1 mm of tape.
Hence, the length L2 of a block, excluding the IBG, is as follows:
L2 = 8,000 (bytes/block)/64 (bytes/mm) = 125 (mm/block).
Therefore, the block length L3, including the IBG, is as follows:
L3 = 125 (mm/block) +15 (mm/block) = 140 (mm/block).
80 bytes20
…
IBG
Record 1
Record 2
15mm
Record 3
…
Record n
IBG
…
L1=8,000 bytes, L2=125mm
L3=140mm
Number of records that can be stored on one magnetic tape
Let us now calculate precisely how many records can be stored on one magnetic tape.
(1) Calculating the number of blocks that can be stored on one tape
Since the length of the tape is 730 m (730 * 103 mm), the number B1 of blocks that can be recorded on one tape is as follows:
B1 = (Length of the tape) / (Length of a block)
= 730 * 103 / 140
19
(FAQ) On each exam, there is at least one question dealing with the calculation of the capacity or performance of a magnetic tape or a hard disk. If you keep these ideas organized in your mind, you can answer these questions because the difference is only in the numerical values.
20
(Hints & Tips) Besides “bytes/mm,” the record density can be represented in “columns per mm” or “bpi.” A column is the same as a byte. “bpi” stands for “bytes per inch,” which is the number of columns per inch. Converting from inches to mm is necessary here, but you need not remember the formula since the relationship between inches and mm will be given in the question.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
71
2. Computer Systems
= 5,214.285…≒5,214 (blocks).21
The fractional part 0.285 is less than one block, so we truncate it.
(2) Calculating the number of records that can be stored on one tape
Since one tape can record 5,214 blocks, and each block has 100 records, the number B2 of records that can be stored on one tape is as follows:
B2 = (number of blocks that can be stored on one tape) * (blocking factor)
= 5,214 * 100
= 521,400 (records)
Performance Calculation
The running speed of the tape is constant when reading or writing data. In theory, the tape begins to accelerate in the middle of IBG and starts to read and write when a constant speed is achieved. When another IBG is found, the tape decelerates and stops in the middle of IBG.
IBG
block
IBG
block
IBG
Running speed of the tape
Time
Stop
Start
Running at a constant speed
Constant Deceleration speed Assuming that the specifications of a magnetic tape are given below, let us calculate the time it takes the magnetic tape to read one block.22
[Specifications of a magnetic tape]
Data transfer speed
Record length
Blocking factor
Start or stop time
320 Kbytes/sec
80 bytes
100
6 milliseconds
Transfer time for one block
Transfer time for one block can be obtained by dividing the length of a block by the data transfer speed:
Data transfer time for one block = (block length) / (data transfer speed)
Block length = (record length) * (blocking factor)
= 80 * 100
= 8,000 (bytes)
Data transfer time for one block = 8,000 (bytes) / 320 (Kbytes/sec)
21
(Hints & Tips) The fractional part of the number of blocks is discarded here, but it actually becomes a short block, which is a block with fewer records than the other blocks.
22
(Hints & Tips) In performance calculations, data is transferred in blocks, so we do not need to consider the length of
IBG.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
72
2. Computer Systems
= 8,000 / (320 * 103)
= 25 * 10-3 (seconds) = 25 (milliseconds)
Time required to read one block
We add the start-up time to the data transfer time for one block.
Time required to read one block
= (start-up time) + (data transfer time for one block)
= 6 (milliseconds) + 25 (milliseconds) = 31 (milliseconds)
Note that the start-up time is added but not the stop time. When being read, no data is transferred until the beginning of a block is reached. Hence, the time until this is achieved is part of the waiting time. After that, data is transferred, but when the data transfer is completed, the stop operation and the program processing are performed concurrently. Thus there is no need to add the stop time.23
2.1.5 Hard Disks
Points
¾
¾
A hard disk is configured with cylinders and tracks.
Hard disks of the sector type do not have IBGs.
A hard disk is a medium that achieves random and high-speed reading and writing of data, consisting of 1 to 10 round disks coated with a magnetic substance on the front and the back sides and rotated at a high speed. If there is only one disk, it is called a floppy disk.24
High-speed rotation
Some disks do not record data on the very top and very bottom sides (protective sides).
Tracks
(Circumference portion)
Read/write head
(portion that reads and writes data)
(as many as the recording sides of the disk)
Cylinders
(Cylindrical portion)
Access arm
23
(Hints & Tips) That stop time is not included is a standard assumption in exam questions.
(Note) Another medium that, like a floppy disk, can read and write and can easily be carried around is MO (magneto optical disk), which is very popular because its capacity is about 600 times that of a floppy disk.
24
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
73
2. Computer Systems
Capacity Calculation
Just as on a magnetic tape, data is recorded in blocks on a hard disk. However, if a block cannot fit into a track, it cannot be recorded.
Block 1
IBG
Block 2
IBG
…
IBG
block n
Block
Cannot be recorded
Track length
First, data is recorded on a track, and when that track is filled, the recording proceeds to the next track which is a track on the corresponding circumference of the next surface. In other words, data is recorded in cylinder units.
Assuming that the specifications of a hard disk are given below, let us calculate actually how many cylinders are necessary to write 100,000 records.
[File specifications]
Record length
Blocking factor
250 bytes
8
[Specifications of a hard disk]
Number of cylinders per disk
Number of tracks per cylinder
Number of bytes per track
Block gap
400
19
13,000
135 bytes
Block length (B)
B = (record length) * (blocking factor) + (block gap)
= 250 * 8 + 135 = 2,135 (bytes)
Number of blocks that can be recorded on one track (N)
N = track length / 2,135 = 6.008…≒6 (blocks) (truncated)
We truncate the answer because a fractional part does not constitute a block.
Number of records that can be recorded on one track (Rt)
Rt
= N * (blocking factor)
= 6 * 8 = 48 (records)
Number of records that can be recorded on one cylinder (Rs)
Rs
= Rt * (number of tracks per cylinder)
= 48 * 19 = 912 (records)
Number of cylinders required to write 100,000 records (S)
S
= (number of records) / Rs
= 100,000 / 912 = 109.649… ≒110 (cylinders) (rounded up).
In general, files are secured in cylinder units, so if there are unused tracks on a cylinder, the
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
74
2. Computer Systems
number of cylinders is rounded up to the next integer.25
Performance Calculation
The access time of a hard disk is calculated as follows:
Access time = waiting time + data transfer time
= (seek time + latency time) + data transfer time
Translator’s note: The waiting time (seek time + latency time) is often called the access time.
The seek time is the time during which the read/write head moves to the track where the data is recorded. The latency time is the time until the desired data come under the read/write head.
The seek time and the latency time are determined by where the head is located, so we use the average values. In actual exam questions, the seek time is always given, and the latency delay can be calculated by the duration of one rotation, which is obtained by the inverse of the number of rotations per time unit. Since the minimum latency time is 0 and the maximum is 1 rotation time, the average latency time is the duration of a half rotation.
Assuming that the specifications of a hard disk are given below, let us calculate actually the access time for reading data contained in one block (5,000 bytes).
[Specifications of a hard disk]26
Number of rotations of the hard disk:
Memory capacity per track:
Average seek time:
2,500 rotations per minute
20,000 bytes
25 milliseconds
Calculation of the average latency time
The fact that the number of rotations of this hard disk is 2,500 rotations per minute means that the disk makes 2,500 rotations every minute. Hence, the rotation time is as follows:
Rotation time = (1 * 60,000 msec/min) / 2,500 revolutions/min = 24 msec/revolution
Note carefully these units. The rotation speed is given in rotations per minute, but the rotation time is in milliseconds. Hence, we need to convert minutes to milliseconds (1 minute = 60 seconds = 60,000 milliseconds).
Since the average latency time is ½ of the rotation time, it is 12 milliseconds.
Calculation of data transfer speed
Since one rotation allows the transfer of data contained on one track, 20,000 bytes are transferred in 24 milliseconds. Hence, 20,000 / 24 (bytes/msec) is the data transfer speed. We can calculate this quotient. But, since it is indivisible, we shall leave it as it is here.
25
(FAQ) On each exam, there is at least one question dealing with the calculation of the capacity or performance of a magnetic tape or a hard disk. If you keep these ideas organized in your mind, you can answer these questions because the difference is only in the numerical values.
26
Seek time/Latency time: The seek time is sometimes called the positioning time. The latency time is sometimes called the search time.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
75
2. Computer Systems
Calculation of data transfer time
The time required to transfer 5,000 bytes is then calculated as follows:
Data transfer time = (amount of data transfer) / (data transfer speed)
= 5,000 / (20,000 / 24)
= 5,000 / 20,000 * 24
= 6 (msec)
Calculation of access time
Access time = 25 msec + 12 msec + 6 msec = 43 msec
The calculations above are based on the rotation speed of the hard disk. However, auxiliary memory, such as a hard disk and a magnetic tape, exchanges data with the computer through input/output channels. 27 It is therefore necessary to install input/output channels with appropriate transfer speeds.28
Capacity Calculation of Sector-based Hard Disk
The term sector refers to the way a magnetic medium is partitioned on floppy disks or hard disks. A sector is an arc of a fan-shaped portion of the disk formed by radial lines drawn from the center of a track in equal intervals. Input/output of a sector-based recording medium is done in sector units, without using IBGs. Each sector is filled with as many records as possible, and the remaining portion of the sector is not used.
Track (on circumference)
Sector (arc)
For example, suppose that each track consists of 12 sectors, each of which consists of 1,200 bytes on a hard disk. To store files whose record length is 900 bytes, there is no sector remainder, as shown below, if the flocking factor is 4.
1,200
3,600
1,200
1,200
Sector
Sector
Sector
Record
Record
900
900
Record
Record
900
900
← 3 sectors are used
← 4 records in a block
3,600
27
Input/output channel: Data-transfer path for exchanging data between auxiliary memory and the computer
(Hints & Tips) The data transfer speed of a hard disk is determined by the rotational speed of the disk, so it is meaningless to have a high-speed channel. Channels are to be selected according to the transfer speed of the disk.
28
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
76
2. Computer Systems
2.1.6 Terms Related to Performance/ RAID
¾
Points
¾
Regarding storage media, there are some terms such as access time, waiting time, transfer time, seek time, and latency time.
RAID is a disk array system for achieving enhanced reliability and/or increased processing speed.
There are some terms related to the performance of storage media, including access time, waiting time, transfer time, seek time, and latency time. RAID, sometimes called a disk array, is a way to control multiple disks placed in parallel as if they were one unit.
Terms Related to Performance
The figure below shows how the various processing times are related to one another, beginning at time S, when a processing unit requests input/output to a hard disk unit, to time T, when the data delivery is completed.
Access time
Waiting time
Transfer time
Seek time
Latency time
S
I/O request
T
Delivery complete
Since the head moves to the desired track while the disk is rotating, the rotation time and moving time partially overlap.29 However, on IT exams, this overlap is almost always ignored.
Hence, there is no problem defining waiting time as follows:
Waiting time = Seek time + Latency time (or Search time)
RAID (Redundant Array of Independent Disks)
RAID describes auxiliary storage in which multiple hard disks are placed in parallel and are controlled as if they were one disk unit so that the input/output speed can be improved and/or reliability can be enhanced. Sometimes the term RAID refers to such an auxiliary storage or a method. It is an attempt to speed up the process by spreading the blocks over multiple disks and reading the blocks simultaneously.
29
(Hints & Tips) A hard disk is rotating as the head approaches the track, so the seek time and the latency time overlap partially. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
77
2. Computer Systems
Main memory
←Block
←Block
←Block
←Block
Disk 1
Disk 2
Disk 3
Disk 4
RAID has 6 levels as described below, from RAID0 to RAID5.
RAID0
This is the method of writing blocks of a fixed size on multiple disks. Access is not centralized on one single unit, so the input/output time can be reduced.30
RAID1
By recording the same data on two disks, this configuration enhances the safety of the data.31
RAID2, RAID3, and RAID4
These are configurations where, in addition to data recorded on the hard disk, there is a disk designated as the error-checking disk to prevent failures. RAID 2 can correct errors. RAID3 and RAID4 can detect errors while they cannot correct errors. In RAID3, data is partitioned in bits or bytes whereas RAID4 partitions data in block units.32
RAID5
This is where each data block is assigned a parity value. Data and parity are written on separate disks, and a failure on a single disk can be recoverable.
Below is an example of RAID5. Here, 4 disks are handled as one group.
Disk 1
Block 1
Block 4
Block 7
Parity 10~12
Disk 2
Block 2
Block 5
Parity 7~9
Block 10
Disk 3
Block 3
Parity 4~6
Block 8
Block 11
Disk 4
Parity33 1~3
Block 6
Block 9
Block 12
Here, the data is divided up into blocks of a certain length, and three blocks are considered to form a unit. For example, Blocks 1 through 3 are a unit, and for each bit, the exclusive logical sum of blocks 1 through 3 is written on a separate disk as parity value 1 through 3. Similarly, the exclusive logical sum of blocks 4 through 6 is written on a separate disk as parity value 4 through 6.
There is also RAID6, in which the parity values are separated as in RAID5 and the data is recoverable even when two disks fail.34
30
(Hints & Tips) The only feature about RAID0 is that I/O is dispersed, so this is not a measure to improve reliability.
(Note) RAID1 is called mirroring since the same data is recorded on separate hard disks.
32
(Note) There is also RAID0+1, a combination of RAID0 and RAID1, already in use.
33
(Note) The parity of RAID5 uses the exclusive logical sum of multiple blocks. Hence, even if one of the disks should fail, the damaged data can be recovered by taking the exclusive logical sum of the other blocks.
34
(FAQ) Often on the exam, there are questions of the form: “Which of the following statements is appropriate concerning
RAID…?” Remember the difference between RAID0 and RAID1.
31
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
78
2. Computer Systems
2.1.7 Auxiliary Storage/Input and Output Units
¾
Points
¾
Auxiliary storage includes hard disks, magnetic tapes, magneto optical disks, CDs, DVDs, etc.
Input/output units include keyboards, image scanners, tablets, displays, printers, etc.
Any storage excluding the main memory is called auxiliary storage. Auxiliary storage can be used to compensate for the insufficient capacity of the main memory. In general, auxiliary storage has larger capacities in comparison with the main memory.
Input and output units include both input units, where data is entered into the computer, and output units, where data is taken out of the computer. A unit equipped with both the input and output functions is called an input/output unit.
Auxiliary Storage
Typical auxiliary storage includes the following media. In the past there was a time when magnetic tapes and floppy disks were the mainstream media; however, recently the main types have been hard disks, magnetic optical disks, CDs, and DVDs.35
Medium
Hard disk
DVD
DVD-ROM
DVD-R
DVD+R
DVD-RAM
DVD-RW
DVD+RW
CD
CD-ROM
CD-R
CD-RW
Capacity small to 300GB
4.7 to 9.4BG
3.95 to 7.9GB
Re-writing
Yes
No
Yes, only once
3.95 to 7.9GB
Yes, many times
700MB
700MB
700MB
No
Yes, only once
Yes
Magneto optical disk36
(MO)
Floppy disk
Hard disk
DAT37
128, 230,
640MB, 1.3GB
1.4MB
a few GB
Max. about
24GB
Yes
Yes
Yes
Yes
35
Properties, etc.
Mostly built-in
Replacement for CD-ROM
Playable on DVD-ROM units Multiple specifications
For software distribution, etc.
For backup, etc.
Requires a dedicated drive for re-writing
Written with magnetism and light and read with light
Good portability
For backup
For backup
(Note) DVDs are optical disks just as CDs are, but with reduced laser-light wavelength, the DVDs have larger capacities.
The record density on DVD is also larger.
36
(Hints & Tips) A magneto optical disk uses light and magnetism for writing data but uses only light for reading data.
37
(Note) DAT is a unit that records audio onto a magnetic tape using digital signals. It was originally designed for music, but it is now used as a backup system because of its low cost.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
79
2. Computer Systems
Input and Output Units
Input units include keyboards, image scanners, tablets, pointing devices, etc. Output units include displays, printers, etc.
Input units
The most common input devices are keyboards and pointing devices. A keyboard is used to enter numerals and characters while pointing devices38 are used to enter coordinate values.
Other input units are shown in the following table.
Unit
OCR
OMR
Mouse
Tablet
Barcode reader Image scanner
Functions, etc.
This reads handwritten characters and printed characters optically.
This reads handwritten marks optically.
This is used to enter coordinate positions of the mouse pointer.
Dedicated devices such as a light pen are used to enter coordinate positions. This reads barcodes (not the numbers printed).
This reads image data such as pictures and photos, converting them to digital data.
Output units
The most common output devices are displays and printers.39
Unit
Descriptions, etc.
This uses a cathode-ray tube; it is inexpensive and
Display CRT comes with a large screen.
This uses liquid crystal; it is expensive but thin and
Liquid crystal40 saves footprint.
The principle is the same as for a copy machine; the
Printer
Laser printer print quality is good, but it is expensive.
Printing is by injection of ink; it is small and
Inkjet printer inexpensive. The printing quality is good but requires specific paper.
Thermal printer
Heat melts the ink, resulting in high-quality, but the
Thermal-transfer
operating cost is high. printer This is noisy but inexpensive; this has duplicating
Dot impact printer capability. Printer for design drawings
Plotter
38
(Note) Pointing device: A pointing device is any unit designed to enter coordinate positions such as a mouse, a tablet, etc. Other examples include trackballs, digitizers, touch screens, etc.
39
(Note) OLED: It is a display using the organic light emitting diode technology. It uses organic materials that emit light when voltage is applied. Compared with LCDs, the viewing angle is larger, the contrast is better, and the response speed is higher, in addition to being thinner and lighter.
40
(Note) Liquid crystal display (LCD): A liquid crystal display can be of various types such as TFT and STN (currently the mainstream is DSTN). STN has a simple structure with low manufacturing costs, but its resolution and contrast are also low. DSTN is an improved version of STN, where contrast is enhanced. TFT has contrast and resolution equivalent to those of CRT but is expensive.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
80
2. Computer Systems
2.1.8 Input and Output Interfaces
Points
¾
¾
I/O interfaces include SCSI, USB, etc.
USB is equipped with a hot plug function and a plug-and-play function. Input and output interfaces are interfaces for connecting peripheral devices such as printers and hard disk units to the PC and for transferring data. Depending on their types, the transfer may be either serial data transfer or parallel data transfer.41
Data Transfer Methods
0
1
0
1
0
1
0
1
Peripheral devices
Computer
There are two data transfer methods between the computer and its peripheral devices: serial transfer and parallel transfer. Serial transfer is the type of transfer in which data output from the computer are serially transferred, one bit at a time. Parallel transfer is the type of transfer in which data bits are transferred in parallel from the computer. For example, if the transfer is
8-bit parallel, there are 8 signal lines:
Peripheral devices
Computer
Serial transfer
Parallel transfer
41
(FAQ) Frequently the exams have questions concerning combinations of I/O interfaces and data transfer methods. It is good to know common I/O interfaces and data transfer methods.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
81
2. Computer Systems
Types of Input and Output Interfaces
Below is a table that summarizes I/O interfaces commonly used today. USB is the most commonly used interface now, equipped with a plug-and-play function42 and a hot plug function.43 Type
RS-232C
Transfer method
Serial
SCSI
Parallel
Centronics
Parallel
GPIB
Parallel
USB
IEEE1394
44
Serial
Serial
Properties, etc.
Originally used to connect a PC to a modem; currently used also for I/O units
Daisy chain allows connection of up to 7 relatively high speed units
Connecting a PC to a printer; high speed but not for long distance Connecting a PC to a measuring device; also known as
IEEE-488
Connecting up to 127 units in a tree configuration
Connecting up to 63 units in daisy chain or tree configuration Daisy chain
This is the connection method used in SCSI and GPIB, where peripheral units are connected along a line. The last unit in the line requires a termination resistor called a terminator.
External hard disk
MO drive
CD drive
Terminator
(to be placed at the end)
USB
USB has two modes: the full speed mode of 12Mbps and the low speed mode of 1.5Mbps. In the full speed mode, relatively high speed units such as printers and scanners are connected. In the low speed mode, relatively slow units such as keyboards and mice are connected. Currently,
USB 2.0 has increased its high speed mode up to 480Mbps, so most peripheral units can be connected. Other interfaces
In addition to the above, there are other interfaces such as IDE (connecting a hard disk), ATA
(IDE standardized by ANSI), and ATAPI (connecting ATA with units other than a hard disk, such as CD-ROM drive and tape streamer). However, currently CD-R and CD-RW units are commonly connected via SCSI and USB.
42
Plug-and-play (Plug and play): This refers to the function of automatically installing and setting the device driver when the peripheral unit or extension card is connected to the computer. The OS checks all the units connected to the computer when it is started up, installing the required device drivers. If the OS does not have the device driver of the unit in its own library, it requests installation of the device driver and, if necessary, even re-starts the computer automatically.
43
Hot plug: This is the function that enables the plug-and-play function while the computer and peripheral unit power is on. 44
IEEE 1394: The standard where the transfer speed is 100Mbps, 200Mbps, or 400Mbps. This is equipped with a hot swap function (peripheral units can be connected or disconnected without having to turn the power off).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
82
2. Computer Systems
Quiz
Q1
Compare DRAM and SRAM:
Item for comparison
Refresh
Level of integration
Access speed
Unit price per bit
Usage
Q2
DRAM
SRAM
Among the components that compose a computer, which ones compose the central processing unit (CPU)?
Control unit
Control flow
Data flow
Operation unit
Input unit
Memory
Main memory
Output unit
Auxiliary memory Q3
Explain the role of cache memory.
Q4
Describe the characteristics of RAID.
Q5
Is USB a serial interface or a parallel interface?
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
83
2. Computer Systems
A1
Item for comparison
Refresh
Level of integration
Access speed
Unit price per bit
Usage
DRAM
Required
High (large capacity)
Slow
Inexpensive
Main memory
SRAM
Not required
Low (small capacity)
Fast
Expensive
Cache memory
A2
Collective term for both control unit and operation unit
A3
It is high-speed small-capacity memory placed between the CPU or the register and the main memory. The main memory is slower than the CPU or the register, so the
CPU processes can be made more efficient by storing frequently accessed data and programs of the main memory in the cache memory.
A4
This is auxiliary memory in which multiple hard disks are placed in parallel and are controlled as if they were one disk unit so that the input/output speed can be improved and/or reliability can be enhanced. Sometimes the term RAID refers to such a method instead of the storage. It is an attempt to speed up the process by spreading records over multiple disks and reading the distributed records simultaneously.
A5
Serial interface (“USB” stands for “Universal Serial Bus.”)
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
84
2. Computer Systems
2.2 Operating Systems
Introduction
Computers do not operate only with hardware. They function only with the use of software called an operating system (OS).
2.2.1 Configuration and Objectives of OS
Points
¾
¾
There is a broadly defined OS and more narrowly defined OS.
The objective of an OS is the effective use of computers.
The definition of an operating system (OS) is not clear. The basic software is called an OS in a broad sense while the control program is called an OS in a narrow sense.
Configuration of OS
An operating system is the basic software that comprehensively controls and manages the entire operation of hardware and software of a computer system. A program referred to as the basic software and its role are shown below. 45
Basic software
Control program (OS in a narrow sense)
- Job management - Data management - Operation management
- Communications management - Task management
- Memory management - Fault management - Miscellaneous
Service program
- Linkage editor - Sort/merge program - System generator - Text editor
- Media conversion program - Debugger -Miscellaneous
Language processor
- Compiler
- Assembler
- Interpreter
- Generator
- Miscellaneous
Service programs and language processors are sometimes called processing programs, which run on the control program. For this reason, the control program is called an OS in a narrow sense.
45
(Note) Operating systems for personal computers include Windows XP, Mac OS X, and OS/2. Those for workstations include Windows Server 2003, UNIX, and Mac OS X Server. For general-purpose machines, there is also MVS developed by IBM. In addition, there is a free OS program called Linux, which is compatible with UNIX.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
85
2. Computer Systems
Objectives of OS
Objectives of OS
An OS attempts to improve the productivity of the entire system by eliminating unnecessary operations and waste of various resources surrounding the computer and by operating the computer system efficiently. The objectives of an OS are organized in the following figure.
Effective use of hardware resources:
Multiprogramming,46 spooling function,47 etc.
Response to various processing modes:
Batch processing, online real-time processing, etc.
Securing reliability and safety:
Improvement in RASIS, etc.
Load reduction of application software:
Virtual memory,48 library management,49 etc.
Support of computer control and operation:
Continuous processing, recording the operation data, etc.
Let us take a look at these individual objectives in detail:
Effective use of hardware resources
Hardware resources include the central processing unit (CPU), memory, I/O units (including channels). etc. The OS controls these resources so that they can be used efficiently.
Response to various processing modes
One computer can handle various processing modes such as batch processing, remote batch processing, online processing, real-time processing, and interactive mode processing. In particular, since online processing has become widespread, the scope of computer applications has been dramatically enlarged.
Securing reliability and safety
Indexes for reliability and safety include RASIS. This is a term coined by taking the initial letters of the words Reliability, Availability, Serviceability, Integrity, and Security.
Load reduction of application software
Application software refers to a program which runs under the control of the OS. The OS provides an environment in which application programs can be efficiently executed.
46
Multiprogramming: It is a mechanism in which programs are processed alternatingly on one CPU so that it can appear as though multiple programs were operating at the same time.
47
Spooling: It is accomplished by using high-speed hard disks as a virtual I/O unit. For example, directly printing on a low-speed printer slows down the processing speed. Instead, the output results can be recorded on a high-speed hard disk first, and then a service program, dedicated only to output, can do the printing when the CPU is not busy.
48
Virtual memory: It is a technique to enlarge the apparent capacity of the main memory so that large-scale programs can be loaded in the memory at a time. Often, auxiliary storage such as a hard disk is used as virtual memory.
49
Library management: It is the function that systematically accumulates primitive programs, object programs, load modules, and other programs developed. This enables integrated management of software assets that are managed individually (by individuals).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
86
2. Computer Systems
Support of computer control and operation
An OS eliminates human intervention as much as possible, as it processes programs (jobs) continuously and records the operation status (log). The record of the operation status is used to check the circumstances under which a fault occurred.50
Human
intervention
Job 1 preparation
Idle time
OS
Resources free
Job 1 execution
Job 1 preparation
Job 2 preparation
Job 1 end
Job 1 execution
Job 2 execution
Job 1 end
Job 2 preparation
Resources free
Automated
Idle time
Job 2 execution
Job 2 end
2.2.2 Job Management
¾
Points
¾
Jobs are units of tasks given to the computer, consisting of multiple programs (job steps).
Job management has functions such as scheduling and spooling.
One of the functions of the control program, the “OS" in a narrow sense, is "job management.” In job management, the priorities of jobs are determined, and the jobs are synchronized. In batch processing, the OS analyzes the contents of JCL (job control language) to assign resources51 and schedule jobs. In interactive mode processing, the OS analyzes instructions entered at the terminal, assigns resources, and performs scheduling. In addition, job management has other functions such as spooling and cataloged procedures.
Job management
Scheduler:
Managing the order of job execution
Master scheduler:
Interface with the operator
Job scheduler:
Managing reception, selection, start, and finish of jobs
Reader:
Reading jobs
Initiator:
Preparing for the beginning of jobs and programs
Terminator:
Spooling:
Clean-up after jobs and programs
Input management for jobs, output management for process results
Cataloged procedure:
Support for execution of typical jobs
50
(FAQ) Concerning operating systems, many exam questions involve knowledge of terms. Be sure to know terms such as multiprogramming, virtual memory, and spooling function.
51
Resource: A resource is a device/unit of various kinds necessary for the computer to operate. It refers to any device related to memory, input, output, control, and other functions; specifically, these include the CPU, main memory, and files.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
87
2. Computer Systems
Scheduler
In job management, jobs are continuously executed under a master scheduler and job scheduler.
The master scheduler plays the role of an interface with the operator via the console panel.52 The job scheduler manages the reception, selection, start, and finish of the jobs.
Reader
This reads the contents of JCL, analyzes them, schedules jobs, and places them in a queue.
Initiator
This selects the programs with high execution priorities among those in the queue and assigns the resources that those programs need.
Terminator
This releases resources that were used by programs just completed. If there is another program following, the terminator starts up the initiator.
Spooling (Spool)
Spooling is the function of the I/O of jobs independent of the programs. Any output results to low-speed units such as a printer are first stored in a spool file. Then, after the program is finished, the output results are printed on the printer from the spool file by the service program of the OS.53
The reason this is done is that, when the I/O unit is slow, directly performing the I/O process would reduce the processing speed of the computer.54
Cataloged Procedures
In job execution directions, typical processing (routine work) such as translation of languages is done in the following way. A set of JCLs is registered together at a separate location, and this registered set of JCLs is called for executing programs. By doing this, the computer prevents JCL errors. This set of registered JCLs is called cataloged procedures.
52
Console panel: It is a unit where the operator interacts with the computer system via key control and monitors its operation and where the system communicates failures, etc. It consists of a keyboard and a display.
53
(Note) For spool files usually a hard disk is used. The service program that performs the sending of spool files to a printer is often called a writer (output writer).
54
(FAQ) Most exam questions on job management are about spooling. They are always in the form of selecting the correct term, so be sure to have accurate understanding of spooling.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
88
2. Computer Systems
2.2.3 Task Management
Points
¾
¾
A task is the smallest execution unit for using a resource.
An interruption takes place in order to switch tasks.
One of the functions of the control program, which is the “OS” in a narrow sense, is “task management.” Task management is the function of controlling the execution of programs and consists of various procedures such as synchronization control of programs, dynamic assignment of resources for program execution, and management of execution priorities of the programs. It also conducts various types of interruption control.
Tasks and Jobs
A unit of processing from the perspective of the user (a single program or a set of programs executed consecutively) is called a job. In contrast, a unit of internal processing executed under the OS is called a task. “Task” refers to a processing unit that subdivides a program process.55
Control of Task Execution
Task management generates tasks required in response to a command and monitors the execution process. When a task generated becomes no longer necessary, the task is eliminated. The state transitions for tasks are shown below:
Running
2. Dispatching
3. Timer interruption Task generation 1. Registration
Ready
5. Input/output interruption 6. Terminate
Task elimination 4. Supervisor call
Waiting
1. A task has been generated, so it is entered into the queue. Æ Move to the ready state.
2. For execution, move to the running state via the task dispatcher. Æ Move to the running state.
3. Time has expired; withdrawn to make way for a task with high priority. Æ Move to the ready state.
4. Withdraw for an I/O instruction. Æ Move to the waiting state.
5. Again waiting to be executed after completion of I/O. Æ Move to the ready state.
6. All processes are now completed. Æ The task is terminated.
Dispatching refers to the step of selecting a task with high priority from among those tasks in the ready state and advancing it to the running state.56 Supervisor call refers to the step of invoking a function of the OS; in state transfer of tasks, this refers to an I/O instruction. I/O interruption is a
55
Process: This term refers to a program being executed and is used interchangeably with the term “task.” “Process” is a word used by some operating systems such as UNIX. In recent years the expression “process” is used frequently.
56
Dispatcher: It is the program of the OS that carries out dispatching; also known as the dispatching routine.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
89
2. Computer Systems
notice that I/O is completed.57
Interruption
Interruption refers to temporary suspension of a program currently being executed for any reason and transferring control to the OS to execute some necessary processing program. There are external interruptions caused by certain specific states of the hardware and internal interruptions caused intentionally when the control program is called from within a program.
The hardware detects interruptions. When the CPU detects an interruption, the OS receives it, changes the program being executed to the necessary state prior to the interruption, examines the cause of the interruption, and transfers control to the corresponding processing routine (program).
The area in which this state is stored is called the PSW (program status word). There is also a possibility that while an interruption process is taking place, another interruption occurs. Priorities are given to these interruptions depending on their types so that multiple interruptions can also be controlled. The table below describes main types and examples of causes for interruptions.
Type of interruption cause Name
Machine check interruption
External interruptions
Clock function interruption
(timer interruption)
Input/output interruption58
Internal interruptions 59
(traps)
External signal interruption
Program interruption
Supervisor calls (instruction interruption) Possible causes
Malfunction of units, fault, power/voltage trouble
Elapsed time of a fixed length (interval timer), reaching a designated time
Input/output completion, input/output unit status change (out of paper, etc.)
Instruction from console panel, external signals
Overflow, underflow, undefined instruction code execution, division by 0, memory protection violation
Input/output operation command, task switching, page fault, control program invoked
2.2.4 Data Management and File Organization
Points
¾
¾
Data management provides integrated methods for accessing files.
File organization includes sequential organization, direct organization, indexed organization, partitioned organization, etc.
Another function of the control program, which is the “OS” in a narrow sense, is “data management.” Data management is the control program that manages data input and output. It provides various file organization methods such as sequential organization, direct organization, and indexed organization. It works as a bridge between logical files processed within a program and physical files whose structures are different.
Data management allows programmers not to worry about the physical structure of the files.
57
(FAQ) State transition of tasks (processes) is almost certain to appear on every exam. Commit the entire figure of state transition to your memory.
58
(Hints & Tips) Issuing of an I/O instruction is notified by supervisor call; I/O completion is notified by I/O interruption.
59
Internal interruptions are intentionally caused by programs, so they are sometimes referred to as traps.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
90
2. Computer Systems
Access Methods
Access methods
Access methods include sequential access, direct access, and dynamic access, which is a combination of the first two.
Sequential access:
Direct access:
Dynamic access:
Processing files sequentially from the beginning60
Processing a specific record directly
Using direct access to find and position a record, followed by sequential access
Sequential access
This is the method of handling records in a file in sequential order from the beginning. This can be performed with almost all recording media. This is suitable for collective processing in which all records in a file are subject to processing.
Direct access
This is the method where a necessary record is directly (randomly) accessed regardless of the order in which the records are stored. This method is used when the file medium is a directly accessible storage medium such as a hard disk. It is used in online real-time systems where only a part of a large number of records stored in a file needs to be accessed for quick update.
Dynamic access
This is the method where direct access is used to find a specific record and then sequential access follows. Similar to direct access, this is used when the file medium is a directly accessible storage medium.
60
(Hints & Tips) Sequential access can be used with most media; however, direct access and dynamic access are limited to directly accessible media such as a hard disk.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
91
2. Computer Systems
File Organization
File organization
File organization methods include sequential organization, direct organization, indexed organization, and partitioned organization,61 etc.
Sequential organization files
Files are processed in the order in which they were recorded on the medium.
Direct organization files
Records are directly processed by key values.
Indexed organization files
Direct access by index and sequential access are possible. Partitioned organization files
One file contains multiple sequential organization files. Sequential organization files
The records in the files are stored in consecutive positions following a certain order. Files of this type can be created on almost all media such as magnetic tape, hard disk, and floppy disk.
In general, only sequential access is possible with these files.
Direct organization files
A storage address on the medium is calculated based on the key value found in each record, and the record is stored in that position.62 To access a record, we first calculate the storage address using the same formula, and the record is read from that location. There is a method where the key value of each record is directly used as the storage address for the record, but this is not very practical, creating a lot of wasted memory if the key values are not consecutive. A more general way is to use a certain type of conversion formula to calculate a storage address from the key value of each record. This is called address conversion (randomization).
Key
Store
Address calculation
Record
Record
Record
Record
Record
Record
…
…
…
Record
Record
Record
Address conversion sometimes produces the same address for different records. In such a case, the record stored first is called the home record while the record assigned to the same address later is called a synonym record.63
61
(FAQ) Questions concerning the characteristics of these organization methods are frequently asked. Know the characteristics of each type of file organization.
62
(Hints & Tips) A special case of direct organization is relative organization, in which the key values are consecutive like
1, 2, 3, … and the key value is itself the storage address of the record.
63
Synonym/Home: Address conversion can take different key values but produce the same storage value. This is called a
“synonym” (word meaning the same thing). If the result of key conversion stores a record, this record is called the home record, and another record that could not be stored there is called a synonym record. A synonym record needs to be stored elsewhere by some other method (e.g. by list).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
92
2. Computer Systems
Indexed organization files
These are files with an index, and they are organized such that the user can access the records by looking up their addresses by index. Sequential access, direct access, and dynamic access are all possible. In each case, the actual records are accessed only after their addresses on the medium are looked up using the index, so not only does the medium contain a basic data area
(prime domain) where the data is stored but also an index area. Further, in order to prevent a situation where records cannot be added to the basic data area, an overflow area64 is also reserved. Partitioned organization files
In these files, sequential organization files are grouped into units called members, each of which is given a name. Then a directory containing these names and their leading addresses is created. Access is allowed to these members. Think of a member as a set of multiple files organized sequentially. Direct access can find the beginning of a member, and sequential access can be used to find a record in that member. Partitioned organization files are used as storage locations for program files and libraries but not very much as data files.
Member A
Member B
Member C
Directory area (Register)
File B
Member area
File A
64
(Note) There are two types of overflow areas: cylinder overflow area and independent overflow area. Overflow records from various tracks are stored in a cylinder overflow area; if a cylinder overflow area becomes full, additional records are stored in an independent overflow area shared by all of the files.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
93
2. Computer Systems
Hierarchical File Systems
UNIX and Windows65 make use of hierarchical file systems as mechanisms to manage files efficiently. Directories and files
A file system has a hierarchical structure consisting of files and directories (directories are file registers). At the top of the hierarchical structure is the root directory, and directories under it are called subdirectories.66
Root
DIR 1
File 0
Root directory
DIR 2
DIR 4
DIR 3
DIR 5
Subdirectory
File 1
File 2
: Directory
: File
File manipulations
When searching for a file, we designate the path showing in which directory the file is located.
There are two methods for doing this. For instance, if the hierarchical structure is as shown in the figure above, we can designate the path in the following ways:67
• Absolute path
Designating a path from the root directory68
<Example> Here is a way to designate “file1.”
\DIR3\file1 (The leading symbol “\” indicates the root directory.)
• Relative path
Designating a path from the current directory69
<Example 1> Here is the designation of file2 when the current directory is “DIR2.”
DIR4\file2
<Example 2> Here is the designation of file2 when the current directory is “DIR4.” file2 65
(Hints & Tips) What UNIX and MS-DOS call “directory” is called “folder” in Windows and MacOS.
(Hints & Tips) Whereas directories and files can be made under a directory, files and directories cannot be made under a file. 67
(FAQ) Concerning hierarchical file systems, there are exam questions like “Choose an appropriate designation as an absolute path or as a relative path.” In those questions, the symbol for separating directories and files, as well as its use, will be explained in the question text.
68
(Hints & Tips) Here we are using the symbol “\” to separate directories and files, but some operating systems use the symbol “/” instead.
69
Current directory: It is a directory in which the user is working at the moment.
66
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
94
2. Computer Systems
2.2.5 Memory Management
¾
Points
¾
Memory management uses two types of memory: real memory and virtual memory.
The basic format of the virtual memory system is the paging method.
Another role of the control program, which is the “OS” in a narrow sense, is “memory management.” Memory management makes the most effective use of the memory as well as compensating for any lack of the main memory capacity. To this end, it effectively uses auxiliary storage as part of the memory.70
Real Memory System71
This is the system that manages the physical space of the main memory. Real memory can be controlled in various ways: partition, swapping, relocation, and overlay methods.
Partitioned method
When a program is placed in the main memory, the main memory is partitioned into several partitions, into which the program is loaded.72 Without memory management, fragmentation occurs, causing a situation which prevents programs from being stored even though empty space exists. Hence, to combine all empty areas together, compaction73 is necessary.
Swapping (roll-in / roll-out)
Swapping refers to execution as the program keeps switching back and forth between the main memory and auxiliary storage. If a program is entered with higher priority than the priority level of the currently executed program, the new program is immediately loaded into the main memory and is executed. However, if there is no space in the main memory, any program in the main memory can be moved to the auxiliary storage. Hence, this system compensates for a lack of main memory capacity by utilizing auxiliary storage. However, if swapping occurs frequently, it means that programs are switched back and forth many times, thus reducing the processing efficiency of the computer system.
Main memory
Program X
Program B
Program A
Swap out (roll out)
Program X
Swap in (roll in)
Program A
70
Auxiliary storage
Memory leak: Sometimes, for some reason, memory in the main memory, secured dynamically by an application, may not get released but remains in the main memory. This is called memory leak. To eliminate memory leak, compaction must be performed.
71
Real memory system or “Real Storage (RS)”: This refers to the actually existing memory; it is the main memory.
72
(Note) In partitioned method, multiple programs can be stored simultaneously, so multitasking is possible.
73
Compaction: It means collecting empty memory areas to form a continuous area; also known as garbage collection.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
95
2. Computer Systems
Relocation (Relocatable)
Relocation refers to the function wherein a program already assigned to a certain area is re-stored in another location. A program whose structure allows it to be relocated is called a relocatable program.74 Overlay method
The physical limitations of the main memory can be eliminated; that is, programs are divided into segment units, and only the necessary segments75 are loaded into the main memory to be executed.
The entire program is stored in auxiliary storage, and the main memory contains only frequently used segments. Exclusive segments, which are never used simultaneously, are loaded from auxiliary storage to the main memory on an as-required basis.
For example, suppose that Segment A is a main routine used with high frequency while Segments
B and C are subroutines called exclusively by Segment A. While Segment B is being executed,
Segment C is in auxiliary storage. When Segment C is called, it is loaded in the area of Segment B.
Consequently, the entire memory capacity of the program is “A + B + C,” but the capacity of the main memory is sufficient if it is at least the greater of “A + B” or “A + C.”
Program
Call
Segment A
Segment B
Segment C
Load
Segment B
Segment C
Capacity actually required
Program capacity Load
(Auxiliary storage)
Segment A
(Main memory)
Virtual Memory
Virtual memory provides a large capacity of storage space regardless of the size of the main memory.76 Programs are stored in virtual memory (normally in auxiliary storage), and only the parts necessary for execution are loaded into the main memory.
Since the program is loaded into virtual memory, the instructions and data is given virtual addresses, which need to be converted to actual addresses (main memory addresses) for the execution of the program. This conversion is implemented by hardware called DAT (Dynamic
Address Translator).
If virtual memory is used, it is necessary to convert virtual memory addresses to physical addresses in the main memory. This address conversion is performed at a high speed by DAT.
Relocation is also performed if the main memory has fragmentation. Since this relocation is done during the execution of a program, it is called dynamic relocation.
We discuss the virtual memory strategies, which include three methods: page, segment, and segment-page. 74
(Note) “Relocatable” means that compaction is possible.
Segment: It is a logical processing unit of a program. Here, we can regard a segment as a subroutine.
76
Virtual memory system or “Virtual Storage (VS)”: It is a conceptual storage that does not actually exist. A program to be executed appears to be loaded into virtual memory, which is a large memory space, while only the portions (pages or segments) of the program with high frequency of use, data, and other parts necessary for the execution get loaded into the main memory.
75
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
96
2. Computer Systems
Page method (Paging method)
In this method, the program is partitioned into units of a fixed size, called pages. A page then becomes the unit for loading into the real memory. Pages are managed by a page table, which has one entry for each page of virtual memory. If the corresponding page is in the real memory, the page fault bit becomes 0. This page fault bit then indicates whether or not the corresponding page is in the real memory.
Page 0
Page n
n
n
0
b
Page 1
⋮
⋮
a n ⋮
Page table
0 a 0
1 b 0
2
1
3
1
⋮
⋮
⋮⋮
⋮
Page n
⋮
Program
Virtual memory
Page 0
Page 1
Page 2
Page 3
Page number
Real memory
Page fault bit
Dynamic relocation
Dynamic address translator
Converts to real addresses
Segment method
In this method, programs and logical sets of data is considered segments. Virtual addresses consist of segment numbers and addresses within the segments. The paging method is only for memory management and as such, the programs do not need to be written with pages in mind. In contrast, in the segment method, in which segments have different capacities, the programs must be written in consideration of the segment sizes.
Segments are logical processing units, so they can be treated as subroutines. However, the flexible lengths are sometimes inconvenient to manage, and the usage efficiency of the main memory may be reduced.
Segment-page method
This is an improved version of the segment method, in which segments are further partitioned into pages. Real addresses are accessed in the order of “segment Æ page Æ relative displacement within the page.”77
77
Relative displacement within the page: It is an address assigned such that the beginning of the page has displacement
0.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
97
2. Computer Systems
Paging Algorithms
If a page necessary for processing is not found in the real memory, an interruption called a page fault occurs, and the page is read into the real memory from virtual memory. This is called page-in. On the other hand, page-out is to move an unnecessary page out to virtual memory.
Page-in and page-out are together called paging.78
If paging occurs frequently, the time for executing the control program increases, reducing the performance. This is called slashing. To minimize the occurrence of slashing as much as possible, various algorithms are proposed to select pages that are subjects of page-outs.
Common page-out methods and their properties are shown below.
Method
LRU (Least Recently Used)
FIFO (First In First Out)
79
Properties, etc.
Pages are compared on the time elapsed since the last referencing. The page with the longest elapsed time is paged out. The page with the longest elapsed time up to the present is paged out.
78
(Hints & Tips) Swapping and paging are similar, but note that swapping takes place in program and segment units whereas paging takes place in units called pages, which are parts of a program.
79
(FAQ) There are exam questions that require specific tracking of page-ins and page-outs. For example, if the order in which pages are used is “1, 3, 2, 3, 5, 2,” and if the main memory has page 3, which page will be the first to be paged out?
For these questions, have clear understanding of ideas like LRU and FIFO.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
98
2. Computer Systems
Quiz
Q1
Explain the roles of task management.
Q2
What type of file organization is this? Sequential organization files are grouped into units called members, each of which is given a name. A directory is created, including these names and leading addresses, and access is allowed to these members.
Q3
Explain swapping.
Q4
What is the unit of loading into the main memory in the page method?
A1
It is the function that controls program execution, performing such functions as synchronized control of programs, dynamic allocation of resources to execute the programs, and managing the executing priorities of the programs. It can also conduct a variety of interruption control.
A2
Partitioned organization
A3
Swapping refers to execution as the program keeps switching back and forth between the main memory and auxiliary storage. If a program is entered with higher priority than the priority level of the currently executed program, the new program is immediately loaded into the main memory and is executed. However, if there is no space in the main memory, any program in the main memory can be moved to the auxiliary storage. Hence, this system compensates for a lack of main-memory capacity by utilizing auxiliary storage.
A4
Page
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
99
2. Computer Systems
2.3 System Configuration Technology
Introduction
Various system configurations are being used to reduce the cost and increase the efficiency of computer systems. These include client/server systems to distribute the load, dual systems to improve reliability, and duplex systems.
2.3.1 Client Server Systems
Points
¾
¾
A client server system is a typical example of distributed processing.
Clients request processing, and servers provide services (processing).
A client/server system (CSS) is a form of computer system where a network is used to distribute processing; it is a type of system configuration.80
Overview of Client Server System
CSS is configured with computers with roles called clients and servers. The servers provide services (processing) such as file management, database management, modification and supply of data, printing control, and communication functions, as requested by clients.
A client sends service requests to a server, receives the results of data processed by the server, and displays the results.
The clients and the servers distribute their processes in an attempt to spread the load of computer processing. In addition, by sharing resources, the user can reduce waste. For instance, by connecting a high-speed printer to a server, the clients can share the high-speed printer. Having one high-speed printer may be less expensive than preparing a low-speed printer for each of the clients, although this depends on the number of clients.
Computers used as a server are generally more high-performance than the client computers. To clarify the functions of various servers, they can be named by their functions, such as file servers, database servers, print servers, and communication servers.81
80
Client/server system: In multiprogramming, if there is one operational computer, the user can run both the client and the server within the one computer. In other words, a client/server process does not mean that everything is distributed. It is one method to achieve distributed processing.
81
(Hints & Tips) If a service request made by a client cannot be provided by the server, that server can become a client and request another server to perform the requested process.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
100
2. Computer Systems
Server
Process request
UNIX, Windows Server 2003, etc.
Process result
Windows XP,
Macintosh, etc.
Client
Types of Server
The table below shows the types of server, depending on the provided functions.82
Type of server
File server
Database server
Print server
Communication server
Services provided
Managing shared files, controlling file access and writing
Managing databases, operating databases with DBMS (Database
Management System)
Managing shared printer, printing when a printing request is received
Providing communication functions with the outside using a network
Characteristics of Client/Server System
Since CSS provides a typical type of distributed processing, the characteristics of distributed processing apply in a straightforward manner.83
Advantages of client/server system
•
•
•
•
•
When the processing is local at clients, the response is faster.
Costs for the entire information system can be reduced, achieving a good cost/performance ratio. Specialized server functions give the system greater economical efficiency and performance.
It is easily extended; clients and servers can be flexibly added as needed.
Even when clients access a server, they are unaware that they are in a distributed environment. Disadvantages of client/server system
•
•
•
The system could be confusing unless the server administrator is clearly identified.
The performance deteriorates if the use gets concentrated on specific servers.
The performance of the entire system depends on the network performance.84
82
(Hints & Tips) Clients and servers do not necessarily have to have the same OS. Where there are multiple servers, they do not have to have the same OS either. In addition, if there are multiple clients, they need not have the same OS either.
83
(Note) If client/server type programs are logically divided into three layers (presentation layer, function (application) layer, and data layer), such a system is called a 3-layer client server system. By distinguishing 3 layers by function, such a system strives to enhance system performance and efficiency for development and maintenance.
84
(FAQ) There have been many exam questions regarding the knowledge of client/server systems. Most of them are about the role of a client or that of a server, so be sure to know these things well.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
101
2. Computer Systems
2.3.2 System Configurations
¾
Points
¾
A dual system is a configuration for high reliability; a duplex system is a configuration for high availability.
There are two types of multiprocessor systems: loosely coupled and tightly coupled systems.
A variety of system configuration patterns are available according to the objectives of information processing. For instance, the reliability of a computer system improves by having multiple units installed. We will look at main system configurations and their characteristics from the viewpoints of reliability, efficiency, costs, and other factors.
Simplex System
This system consists of one CPU only. Reliability and the processing capabilities are inferior in comparison with other configurations, but it is economical. This configuration is commonly used.85
Dual System
This is a system configuration in which two CPUs perform the same processing and compare the processing results to each other. This configuration is applied when the process is not allowed to stop, even for a moment. If one CPU fails, the system cuts off the failed CPU and continues processing on the other CPU. Reliability is extremely high, but this system is expensive.86 87
Comparing
processing results1 85
DCE (Data Circuit-terminating Equipment): This unit converts signals received from communication line, sends them to data terminals, and also executes exactly the opposite operation. Normally, this unit is connected at the end of a communication line and functions as an interface with a computer. A modem (modulator-demodulator) is used on an analog line, and DSU (Digital Service Unit) is used on a digital line.
86
CCU (Communication control unit): This unit controls the reception and transmission of data, performs error control, and assembles and decomposes characters.
87
DISC: Auxiliary storage
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
102
2. Computer Systems
Duplex System (Standby System)
In this system, two CPUs are prepared, where the primary system is used for online processing while the secondary system is used for low-priority processing such as batch processing.88 If the primary system fails, the online processing gets switched to the secondary system.89
Switching time is said to be anywhere from several tens of minutes to several hours. Reliability is lower than that of a dual system, but this system is better from the standpoint of costs, so it is more commonly used than the dual system.
Switching unit
Switching unit
(Primary system)
(Secondary system)
Multiprocessor System (Concurrent Processors)
In this system, multiple CPUs and CCUs are sharing and processing tasks, so it is a system configuration with high processing efficiency. There are two types of multiprocessor system. One is LCMP90 (loosely coupled multiprocessors), in which multiple computer systems are controlled by separate operating systems. The other is TCMP91 (tightly coupled multiprocessors), in which multiple computer systems share the main memory and are controlled by the same operating system. The figure below shows an example of TCMP.
92
88
(FAQ) There are many exam questions on characteristics of dual, duplex, and multiprocessor systems. The key term for each system is as follows: “comparing process results” for dual, “switching the units” for duplex, and “sharing the main memory” for multiprocessor systems.
89
(Note) If a failure occurs in the primary system, it takes time to switch to the secondary system. This is because the batch processing or whatever else is being executed in the secondary system must be suspended, and the OS must be booted for the online system. A hot standby system configuration can solve this by standing by, ready to switch at any time.
In this case, the OS for the online system stays on, so switching can occur immediately.
90
LCMP (Loosely Coupled Multiprocessor): Each CPU has its own main memory and independent OS. CPUs are joined by a high-speed network or shared path. This is a configuration where independent computer systems are connected via a network.
91
TCMP (Tightly Coupled Multiprocessor): One main memory and one OS are shared in this configuration. Each CPU can perform identical processes, so even if one CPU fails, the processing can continue, albeit with lower performance. This configuration is highly reliable and thus is used in systems where a high level of processing capability is required.
92
MM: Main memory
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
103
2. Computer Systems
2.3.3 Centralized Processing and Distributed
Processing
Points
¾
¾
Centralized processing has a high level of safety but is inflexible.
Distributed processing is economical, but its real substance is hard to understand. Depending on how the computers are placed physically, there are two types of processing: centralized processing and distributed processing.
Centralized Processing
Centralized processing is a system configuration in which one computer is connected with many terminals, and the one computer alone does all of the processing. It is easy to maintain the consistency of data, and it is easy to manage the resources. These merits contributed to the popularity of this configuration in which a general-purpose computer is used as the host in centralized processing. Below is a summary of relative comparison with distributed processing.
Advantages of centralized processing
It is easy to improve the cost/performance ratio. (Grosch's Law93)
Operation and maintenance require a smaller staff. Safety level of the system is high.
Disadvantages of centralized processing
Extendability is poor to keep up with new technologies.
Backlog can be easily accumulated.94
Overhead of the OS is significant.
Recovery of a host failure is time-consuming.
A failure has a far-reaching effect.
Distributed Processing
Distributed processing is a system configuration in which multiple computers connected via a network perform the processing. Since the processing is done through a network, the processing time is longer than that of centralized processing. But, the merit is that a failure of one computer does not affect the entire system. Below is a summary of relative comparison with centralized processing.95 Advantages of distributed processing
Management responsibilities are clear. (Management responsibilities can be delegated to each organization.)
Effects of system failure are local.
Maintenance is easy (locally closed).
It is economical as only necessary units are installed
93
Disadvantages of distributed processing
Its real status can be hard to understand.
It is difficult to identify trouble spots.
Network performance has great impact.
Data inconsistency can occur easily.
Individual units are managed carelessly.
Grosch's Law: It states that “performance is directly proportional to the square of the price.” If the price of a computer doubles, the performance quadruples. However, technological advancement has reduced the prices of devices significantly, so this law is no longer applicable.
94
Backlog: It means systems, software, programs, etc. that are necessary to develop but the development of which has not even begun. The term often refers to those that are held back in the IT department within a company.
95
(FAQ) There are exam questions where you are required to identify the characteristics of centralized processing and distributed processing. For example, questions may be of the form “Which of the following is an appropriate characteristic of a centralized processing system?” Know the advantages and disadvantages of each processing type.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
104
2. Computer Systems
As shown in the following table, distributed processing can be classified according to the distribution status of functions and loads. It is said that vertical load distribution does not exist in reality. Configuration
Function
Horizontal distribution
Vertical distribution
Function distribution
Load distribution
Horizontal function distribution
Vertical function distribution
Horizontal load distribution
Horizontal function distribution
A horizontal function distribution system is a system in which computers are classified according to type of application and type of data; examples include processing function distribution and database distribution. For instance, in financial institutions, host computers may be classified into those in an information system and those in an accounting system; this classification is based on the type of processing, so it is an example of processing function distribution. Database distribution means that computers are located in appropriate locations based on the contents of data. Horizontal load distribution
This is a system in which multiple computers perform processes jointly when an application is executed. When a process is requested, an idle computer responds. In this mode, if one computer fails, the process switches to another computer and is continued. Hence, this system is quite effective in time of failure. A tightly coupled multiprocessor system is an example of this type.
Vertical function distribution
This is a system where the processing function is shared among workstations belonging to individual users as well as computers shared by multiple users. Here, there is a vertical relationship in regard to the processing function. A client/server system is a typical example of a vertical function distribution system.96
96
(Hints & Tips) A client/server system appears as if it were horizontal distribution, but it is properly classified under vertical function distribution. Since one server performs processes of multiple clients, there is a vertical relationship in functions. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
105
2. Computer Systems
2.3.4 Classification by Processing Mode
Points
¾
¾
In batch processing, data is stored and processed all at once.
In real-time processing, data is processed at the moment they come into existence.
From the standpoint of processing modes, system configurations can be classified into two categories: batch processing and real-time processing. They can also be classified by whether or not they are connected to a network.
Processing mode
Batch processing97
Real-time processing
Operation mode
Center batch processing
Remote batch processing
Interactive mode processing
Online transaction processing
Real-time control
Connection method
Offline
Online
Batch Processing Systems (One-Time Processing
Systems)
The word “batch” means a “bundle.” Batch processing is any method in which data to be processed by a computer are stored for a certain period of time or are saved on original sheets or storage media and are later processed all at once. Such systems have the following characteristics:
•
•
•
The computer can be used efficiently because the processing is done all at once.
It is suitable for routine and repetitive processing (standard tasks).
The results are not immediately obtained because the processing is collectively done.
In center batch processing systems, the processing takes place at a central computer center; in remote batch processing,98 the batch processing is performed from a terminal at hand via a communication line.
Interactive Processing Systems
This is a method in which the processing is performed in the mode of a dialogue with the computer. We can proceed with the processing while communicating with the computer and can correct errors immediately once they are noticed; therefore, it is a processing mode suitable for program development, etc. An interactive processing system requires prompt responses.
Also, an interactive processing system requires an OS in which the multiprogramming function is
97
Batch processing and real-time processing: Batch processing is where data is stored and processed all at once.
Examples include the calculations of electricity payment, water payment, and gas payment. Payroll calculation is also an example of batch processing. Real-time processing is where data is processed immediately when they are generated.
Seat-reservation systems of airlines and trains are examples of this type.
98
Center batch process and remote batch process: Center batch processing does not use communication lines. Remote batch processing is where a terminal unit at hand is used to perform batch processing at a host computer at a remote location. It is sometimes called RJE, which stands for “Remote Job Entry” as the user enters a job at a remote site.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
106
2. Computer Systems
supported, and the processing is performed using TSS.99
Online Transaction Processing Systems
A transaction is a series of data or instructions, sometimes called a transaction file or transaction data. Online transaction processing refers to the processing of data updates on an online file at the computer center via a connected online terminal. Examples of online transaction processing include savings account systems in banks and train and airline seat-reservation systems.
Online transaction processing requires specialized (dedicated) terminals with a high level of usability aimed at enhancing the processing efficiency, such as bank ATMs and terminals for issuing reserved-seat tickets at reservation windows. In addition, since data is shared by many terminals, there is a risk that simultaneous access to the same data could cause problems such as deadlock100 or data destruction. Hence, attempts are made to enhance the data maintainability.
Real-Time Control
In general, real-time control refers to the method by which data is processed in real time once a processing request is made and the result is immediately reported to the requester. This concept also includes online transaction processing. However, in a narrow sense, this refers more specifically to the processing mode at places like manufacturing plants, where the system is interlinked with a sensor system tracking the physical motions of objects to be controlled, processing corresponding to the external signals are immediately executed, and the results are immediately sent back to units on production lines (e.g. robots) as control signals. Production control systems at steel plants and automobile factories are examples of this type. Another example is a 24-hour monitoring process of electrical system managed by computers; if there is something wrong with the system, the system reports it to the maintenance company in real time, or the unit that has detected it gets cut off in real time.101
99
TSS (Time Sharing System): The system authorizes multiple programs to be executed in a specific order for an extremely short duration at a time (several milliseconds at a time). This procedure is repeated so that the execution of each program can get completed within a certain period of time. From the user's viewpoint, there is no mutual interference, so each user feels as if he or she were the only one exclusively using the computer system.
100
Deadlock: It is a situation where the system gets stuck because multiple tasks (programs) try to access the same resource (file, database, etc.) and go into the waiting mode.
101
(FAQ) Many exam questions involve characteristics of remote batch processing and interactive processing. Both are online, but note that remote batch is a type of batch processing while interactive processing is a type of real-time processing. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
107
2. Computer Systems
Quiz
Q1
Explain the roles of clients and servers in a client/server system.
Q2
What is the type of system configuration shown below?
Switching unit
Switching unit
(Primary system)
(Secondary system)
Q3
As shown below, distributed processing is classified by the distribution status of functions and loads. To which category does a client server system belong?
Configuration
Function
Horizontal distribution
Vertical distribution
Function distribution
Load distribution
Horizontal function distribution
Vertical function distribution
Horizontal load distribution
Q4
What is the difference between batch processing and real-time processing?
A1
Clients: They send service requests to servers.
Servers: They perform the requests sent by clients and return the processing results to them. A2
Duplex system: Distinguishable points are the existence of two CPUs and the switching units.
A3
Vertical function distribution
A3
Batch processing: Data is stored and processed all at once.
A4
Real-time processing: Data is processed immediately when they are generated.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
108
2. Computer Systems
2.4 Performance and Reliability of
Systems
Introduction
To evaluate computer systems, various methods are available. While good performance (high processing speed) is important, fault-tolerance (high reliability) is also significant.
2.4.1 Performance Indexes
¾
Points
¾
Response time, throughput, and turn-around time are used for performance evaluation.
Instruction mix and benchmark are used for performance indexes.
To evaluate the comprehensive performance of computer systems, including their software and hardware, we can use various criteria such as response time, throughput, and turn-around time.
Indexes to evaluate performance, especially the hardware, include instruction mix and benchmark.
Terms Related to Performance Evaluation of Computer
Systems
In evaluating the performance of computer systems, it is important to calculate the processing time of jobs and programs.102
Response time
This is the amount of time between the completion of input at an input unit and the beginning of output at an output unit. For example, when a processing request is made at the keyboard of the computer, this time refers to the amount of time it takes until the result is shown on the display unit or until the printing begins. This is mainly used to evaluate the performance of an online system. Throughput (Processing capability)
This refers to the amount/number of jobs that can be processed by the computer system within a certain unit of time, or the amount of time required to process a certain job. This processing time includes the exclusive CPU time and process-waiting time such as preparation for I/O operation and clean-up time.
102
(Hints & Tips) There are instruction mixes and benchmarks for evaluating the performance of computers, and instruction mixes are for hardware evaluation. However, even if hardware is very fast, the entire system performance becomes poor if the performance of the OS is poor. Hence, the performance of hardware is often used only for reference.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
109
2. Computer Systems
Turn-around time (TAT)
Technically, this refers to the amount of time it takes information to make the rounds of the system.
In batch processing, this is the duration between submission of a program at the window and the time when the results are obtained. In business operations, this is the duration from the time when a client places an order to the time when the ordered product is shipped and reaches the client.
Instruction Mix103
An instruction mix is used to compare the performance of hardware in computer systems. Even if the hardware is fast, if the performance of the OS is poor, the performance of the entire system becomes inferior. An instruction mix is to use an average program and calculate the average instruction execution time per instruction and MIPS value,104 based on the execution frequency of each instruction.
Under these conditions, let us do some specific calculations of the MIPS value.
Instruction group
A
B
C
Execution speed (microsecond)
0.1
0.2
0.5
Frequency of appearance
40%
30%
30%
First, let us calculate the average instruction execution time. The execution speed of each instruction is expressed in microseconds (10-6). The average instruction execution time is the sum (over all instructions) of the products of the execution time of instructions and their respective frequencies. Average instruction execution time = 0.1 * 10-6 * 0.4 + 0.2 * 10-6 * 0.3+0.5 * 10-6 * 0.3
= (0.04 + 0.06 + 0.15) * 10-6
= 0.25 * 10-6 (seconds/instruction)
The average number of instructions executed per second is the inverse of the average instruction execution time, so it is obtained as follows:
Average number of instructions executed per second = 1 / (0.25 * 10-6)
= 4 * 106 (instructions/second)
= 4 (MIPS).105 106
FLOPS107 is used as an index to evaluate the performance of floating-point operations.
103
(Note) An instruction mix for scientific calculations is called “Gibson mix,” and one for business calculations is called
“commercial mix.”
104
MIPS (Million Instructions Per Second): This is the performance index expressing the number of machine instructions, in millions (106), that can be executed per second. This is just for the performance of hardware, so again it is used only for reference.
105
(FAQ) Exams do have questions where you are asked to calculate MIPS values given an instruction mix or to calculate the average clock count per instruction. You would want to be familiar with these calculation questions.
106
Clock: This refers to the frequency of a clock signal generated by a circuit called a clock generator. Since instructions inside the CPU are synchronized to this clock signal as they are executed, the higher the clock frequency is, the more instructions can be executed in a given period of time. For example, if the clock frequency is 200MHz, there are 200 * 106 clock signals per second. In general, one instruction takes several clocks.
107
FLOPS: It stands for floating-point operations per second. This is an index expressing the number of floating-point operation instructions executed per second. If it is expressed in millions (106), it is called MFLOPS.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
110
2. Computer Systems
Benchmark
A benchmark is used to compare and evaluate the comprehensive performance of computers, including the hardware and the OS, by measuring the standard program execution time.108
2.4.2 Reliability
¾
¾
Points
Reliability indexes include RAS, RASIS, and the bathtub curve.
Points to be aware of in reliability design are fail-safe and fail-soft.
The level of reliability required for information systems varies depending on the purpose for which the systems are used. Sometimes the economical factor must be sacrificed to achieve a high level of reliability. In some other situations, not only the subject of reliability is focused on the operation of the system but also the information handled by the system needs to be reliable as well.
Reliability Indexes
Reliability is the degree to which system operation is stable. The ideal case is that the system does not fail, but there is no system that does not ever fail.
RAS/RASIS
Both of the terms RAS and RASIS are acronyms of elements that help computer systems to operate in a stable manner. RAS stands for the first three elements of RASIS:
R
A
S
I
S
R: Reliability (Few faults)
A: Availability (High level of availability109)
S: Serviceability (Easy detection of fault parts)
I: Integrity (Consistency of data)
S: Security (Prevention of invalid access)
108
(Note) An example of a technical calculation benchmark is SPECmark, and an example of a transaction benchmark is
TPC. TPC-C is the frequently used benchmark under TPC which directly responds to actual business applications.
109
Availability: Availability refers to the probability that the system is maintaining its functions (operating) at any given time or the percentage of the duration when the functions are maintained during a certain period of time.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
111
2. Computer Systems
Bathtub curve
Failure rate →
The bathtub curve is used to illustrate the concept of hardware lifecycle. Hardware may fail during the initial period of its operation due to defective parts, etc., but the probability at which these failures occur decreases gradually as repairs and replacements are made. After that, because of wear and tear of various parts, the probability of failures increases, and eventually its life is determined to be over. This curve is shown below.110
Early failure period111
Stable failure period112
Wear-out failure period113
Time →
Reliability Design Points
A highly reliable system that can continue to operate even when some part of the system fails is called a fault-tolerant system. Common technologies for configurations of highly reliable systems include fail-soft, the function enabling the system to continue its operation, perhaps with lower performance or fewer functions, when a failure occurs, and fail-safe, the function enabling the system to operate safely by avoiding risky conditions when a failure occurs.
Fail-soft
This refers to the function in which, when a failure occurs, the failed part gets cut off and the system continues to operate, perhaps with a lower performance level (fall back114). In a duplex system, normally the two systems are independently processing data, but if one system should fail, the configuration would switch the processing to the other system and would carry on the processing. In addition, when a failure occurs in multiprocessors, the system continues its services by cutting off the failed processor. This too is a system configuration with fail-soft in mind.
Fail-safe
This refers to the function in which, when a failure occurs, the system locks its functions in a safe mode established in advance to control the extent of the impact of the failure.115 This is just like the measure where all railroad lights turn red when an accident has occurred. In system configurations where two systems compare the processing results of each other, such as in a dual system, when the compared results are different, the system in which a failure is determined to have occurred is cut off while the operation continues on.
110
(Note) The bathtub curve is so named because the graph showing the relationship between the failure probability and time resembles the shape of a bathtub.
111
Early failure period: It is a period of failures at the beginning of unit use. These failures become less frequent as time passes. 112
Stable failure period: The unit is stable during this period, with less frequent failures.
113
Wear-out failure period: A certain period of time has passed, and failures become more frequent during this period.
114
Fall-back: In a fail-soft computer system, processing continues at a lower level of functionality; this is called a fall-back or a fall-back operation.
115
(Hints & Tips) Fail-soft and fail-safe are similar words, so do not confuse them.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
112
2. Computer Systems
Fool-proof
This term refers to a measure that prevents an unintentional use of a program from causing a failure, especially when indefinitely many users use the same program. If one individual is using a particular program, the way the program is written does not create a major problem, but when there are indefinitely many users, how the program gets used is hard to predict.116
2.4.3 Availability
¾
Points
¾
MTBF is the time when the system is operating properly, and
MTTR is the time when it is being repaired.
Availability is the ratio of the time when the system is operating properly. One of the indexes in RASIS is “A” for availability, which means the operation rate. The availability is calculated using MTBF and MTTR as follows:
Availability:
A = (MTBF) / (MTBF + MTTR)
MTBF and MTTR
Suppose that the operation status of a computer system is as shown below:
Time →
Operating
Being repaired T1
D1
Operating
Being repaired Operating
Being repaired T2
D2
T3
D3
…
Operating
Being repaired Tn
Dn
MTBF (Mean Time Between Failures)
This is the average length of time that the system continues to operate without a failure. The larger MTBF is, the more reliable the system is. Therefore, this is used as an index of reliability
(“R” in RASIS).117
MTBF = (T1 + T2 + T3 + … Tn) / n
(Here, “n” is the number of intervals the system was operating without failure.)
116
(FAQ) There are exam questions concerning what each of the letters RASIS means as an index for computer system reliability. At least know what RAS stands for.
117
(Note) Functions that improve MTBF include error detection, automatic 1-bit error correction, instruction re-try, etc.
These are functions that prevent the computer system from coming to a stop. Functions that improve MTTR include log output. By looking up logs, the cause of failure can sometimes be identified. Remote maintenance also helps detect a failure promptly, enhancing MTTR.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
113
2. Computer Systems
MTTR (Mean Time To Repair)
This is the average length of time required for repair when a failure occurs. The shorter the repair time is, the better the system is. Therefore, it is used as an index of serviceability (“S” in
RASIS).
MTTR = (D1 + D2 + D3 + … Dn) / n
(Here, “n” is the number of intervals the system was operating without failure.)
Calculation of Availability
To calculate the availability, the serial connection and parallel connection sections must be calculated differently. The basic ideas are described below.118
Availability in a serial connection system
The availability of an entire serial connection system as shown here is the product of the availabilities of each unit. Here, P1, P2, and P3 are the availabilities of the respective units shown in the figure.
Unit 1
P1
Unit 2
P2
Unit 3
P3
Availability of the entire system = P1 * P2 * P3
Availability in a parallel connection system
Suppose that we have, as shown below, a system in parallel connection where the system operates as long as at least one of Units 1, 2, and 3 is operating. Here, the availability is calculated using the fact that the probability of the entire sample space is 1.
Unit 1
P1
Unit 2
P2
Unit 3
P3
Availability of the entire system
= 1 – (Probability that all units fail simultaneously)
= 1 – (Prob. that Unit 1 fails) * (Prob. that Unit 2 fails) * (Prob. that Unit 3 fails)
= 1 – (1 – P1) * (1 – P2) * (1 – P3)
118
(FAQ) There is always a question involving a calculation of availability. Make sure you understand correctly how to calculate it.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
114
2. Computer Systems
Availability in a system where serial and parallel connections are combined
Suppose that there is a system which operates if Unit 1 is operating AND at least one of Units 2 and 3 is operating. In this case, we consider that Unit 1 and the parallel section (Units 2 and 3) are serially connected.
Unit
P2
Unit
P1
Unit
P3
Availability
= (Availability of Unit 1) * {1 – (Prob. that Units 2 and 3 fail simultaneously)}
= (Availability of Unit 1) * {1 – (Prob. that Unit 2 fails) * (Prob. that Unit 3 fails)}
= P1 * {1 – (1 – P2) * (1 – P3)}
Let us consider more complicated configurations.119 Even though the two systems below may appear similar, the availabilities are different. Here, the letter α in the figure indicates the availability. [Configuration 1] The two parallel sections (inside the dotted lines) are serially connected.
α
α
α
α
Availability = {1 - (1 - α)2} * {1 - (1 - α)2} = α2 (2 - α)2
[Configuration 2] The two serially connected units (inside the dotted lines) are connected in parallel. α
α
α
α
Availability = 1-(1 - α2) * (1 - α2) = α2 (2 - α2)
119
(Hints & Tips) Note that similar configurations have different availabilities.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
115
2. Computer Systems
Quiz
Q1
Explain the meanings of the following terms: “MIPS,” “response time,” “throughput,” and “turn-around time.”
Q2
Explain the meaning of RASIS.
Q3
Explain the meanings of MTBF and MTTR.
Q4
Express the availability using MTBF and MTTR.
Q5
Calculate the availability for the entire system configuration shown below. A and B are units, each of which has an availability of 0.97. The entire system is assumed to be in operation if at least one of the units is operating.
A
B
A1
MIPS:
It is an acronym standing for “million instructions per second.” This indicates the number of instructions that can be executed in one second, expressed in millions (106). It is one of the performance indexes of hardware. Response time:
It is the length of the time interval from data transmission to the return of processing results. It is one of the performance indexes of an online system. Throughput:
It is an evaluation measure of the job-processing capability of a computer system within a given amount of time. It is one of the performance evaluation indexes of a computer system.
Turn-around time: It is the length of the time interval from a job request to the complete output. Mainly a concept used in batch processing, this is an index of system evaluation including its operation.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
116
2. Computer Systems
A2 RASIS is an acronym of elements that help computer systems operate in a stable manner.
R
A
S
I
S
R: Reliability (Few faults)
A: Availability (High level of availability)
S: Serviceability (Easy detection of fault parts)
I: Integrity (Consistency of data)
S: Security (Prevention of invalid access)
A3
MTBF: (Mean Time Between Failures) This is the average length of time that the system continues to operate without a failure.
MTTR: (Mean Time To Repair) This is the average length of time required for repair when a failure occurs.
A4
Availability:
A5
Note that the connection is parallel.
Availability = 1 – (1 – 0.97) * (1 – 0.97)
= 1 – 0.03 * 0.03
= 1 – 0.0009
= 0.9991.
A = (MTBF) / (MTBF + MTTR)
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
117
2. Computer Systems
2.5
System Applications
Introduction
Various systems have been developed using networks and databases. Close to our daily life are the Internet and database services (generally called commercial databases). Examples of applications of multimedia systems include 3D graphics.
2.5.1 Network Applications
¾
Points
¾
Uses of the infrastructure include the Web, the Internet, intranets, and extranets.
Application systems include Internet shopping, groupware, and debit cards.
Today, our information society has networks spanning all over the world like a gigantic web.
Systems using networks themselves and application systems with add-on values are available.
This is a rather new area, so there are not many exam questions on this topic, and they are relatively easy. Most of the questions simply require knowledge of the terms, so be sure to memorize them to improve your exam scores.
Uses of Infrastructure
“Infrastructure” means “foundation” or “basis.” In computer systems, this word refers to the foundation of software and hardware to form the systems. For instance, in network construction, various components such as communication lines, communication units, and the charge system of the communication lines are parts of what is known as the communications infrastructure.
Web (WWW: World Wide Web)
This is the information search system in the hypertext format, developed by researchers at
CERN.120 Since information distributed all over the world is mutually linked by this network using hypertext,121 a name meaning “global spider web” was given to it.
WWW is a mechanism on the server that records information in the form of an Internet homepage.
Software that accesses WWW and displays it on a screen is called a Web browser or simply a
“browser.”
The Internet
120
CERN (Conseil Europeen pour la Recherche Nucleaire) (European Council for Nuclear Research): It is a quantum physics research institute jointly funded and operated by 12 European counties, but generally it is known as the institute which developed WWW on the Internet. Its name has now been changed to Laboratoire Europeen pour la Physique des
Particules, but the abbreviation remains the same.
121
Hypertext: It is a structure in which pointers are placed within texts so that links can be made to jump from those pointers to other texts and pictures. To create a document in the hypertext format, one uses HTML (HyperText Markup
Language). To identify a WWW server address, we can use URL (Uniform Resource Locator).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
118
2. Computer Systems
It is a collection of networks all over the world connected together by TCP/IP. There is no government organization or designated organization managing it in an integrated manner. Instead, the technical support and resource management are done by volunteer organizations.
ARPANET,122 created by the United States Defense Department, set the foundations for the
Internet.
Intranets
An intranet is a company-wide network applying the technology of the Internet. Normally, a firewall123 is set up between the Internet and a company-wide network in order to prevent leakage of confidential information of the company.
Firewall
Internet
Company-wide network Extranets
An extranet is an intranet extended over numerous companies. In general, an extranet is built by connecting intranets to the Internet.
Intranet at
Company A
Intranet at
Company B
The Internet
Mobile communication
Mobile communication124 is an environment in which the network can be accessed from any location. Today, communication is the mainstream, so we can send and receive e-mails on the
Internet and obtain a variety of information all with one telephone.
Satellite communication
Satellite communication is a wireless communications system using a communications satellite.
A broadcasting station transmits (uplink) a huge amount of information to a stationary communications satellite located 36,000 km above the equator (in a stationary orbit), and the information is distributed all at once (downlink) to various receiving stations on the earth. A large amount of information can be transmitted to many points.
CATV
New types of services using CATV (Community Antenna TeleVision, or Cable TV) are being considered and discussed commercially, such as Internet connection, telephone services, experiments involving PHS (Personal Handy-phone System), and VOD (Video On Demand).
CATV is expected to be a major part of the infrastructure in the multimedia era.125
122
ARPANET (Advanced Research Project Agency Network): This is a nationwide computer network developed under the sponsorship of the Advanced Research Project Agency (ARPA) of the United States Department of Defense. It is the predecessor of the Internet.
123
Firewall: It is the mechanism which is located between the Internet and a company-wide intranet to manage data communication and to protect the internal network from external attacks and invalid access. The word could also refer to this functional role.
124
Mobile/mobile computing: “Mobile" refers to any information device that can be carried around, including cell phones, PHSs (personal handy-phone system), and notebook PCs. “"Mobile computing” refers to the mode of using any of these information devices to have access to the company network from the outside.
125
(Note) CATV began as a reparation facility in remote areas and a community facility in rural regions. Today, urban
CATV, which can provide broadcast services on many channels, is getting attention as a new-generation component of the
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
119
2. Computer Systems
Application Systems
A network application system is a social system using a network. Specifically it includes the following: Internet shopping
This is a system in which the user can shop at a virtual store set up on web pages. To make a payment, the shopper can use his or her credit card or go to a nearby convenience store to pay.
Groupware
This is software for communication within an organization or for information sharing. It has functions such as electronic mail, schedule sharing, document sharing, and workflow.
Debit cards
This is a service whereby a cash card issued by a bank can be used to make payments. The money for the payment is directly withdrawn from the bank account in real time.126
infrastructure. Coaxial cables are used for distribution so that high-quality images can be received.
126
Debit cards have been traditionally called bank POS; cash cards are used instead of cash.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
120
2. Computer Systems
2.5.2 Database Applications
Points
¾
¾
An example of a database application is a data warehouse.
Applications in business include corporate accounting, inventory management, document management, and sales support.
One type of database application system is a data warehouse. Application systems in which databases are applied in business include corporate accounting systems, 127 inventory management systems,128 document management systems,129 and sales support systems.130
Data Warehouse
A data warehouse is a company-wide database to support decision-making. The idea is to have a large amount of data stored, organized, and used to help make business decisions. Sometimes it is called an informational database.
Data Mining
This refers to a technology or method of drawing out tendencies, trends, correlations, and patterns necessary for management and marketing, through dialogues with a large amount of raw data.
Whereas a data warehouse normally analyzes various data based on some hypothesis, data mining discovers trends and patterns in order to establish the hypothesis.
Data Mart
A data mart is a database which stores data obtained from a data warehouse. The data stored in a data mart, is selected and summarized according to the purposes of a specific user group.
Whereas a data warehouse contains information for the entire company, a data mart has a relatively small amount of data tailored for the target users.
127
Corporate accounting system: It is a system in which the accounting procedures of a corporation are computerized in an attempt to make the accounting tasks more efficient and quicker and to obtain timely understanding of the business and managerial records.
128
Inventory management system: It is a system to keep the production (purchase) and demand in balance, managing the inventory such as products and raw materials kept by the company at an optimum amount. In a retail store such as a supermarket, the sales information entered at POS terminals is collected and analyzed so that the demands can be predicted and more products are automatically ordered, taking into account safe inventory volumes and optimum amounts to purchase. 129
Document management system: It is a system in which a corporation manages various types of documents and sources; document search is possible from a variety of fields such as the storage location or contents of the document. It is an attempt to make document management and document preparation more efficient, e.g. to avoid duplicate preparation of the same document.
130
Sales support system: It is a system that supports making sales plans and business plans, based on accumulated sales information. FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
121
2. Computer Systems
OLAP
OLAP (OnLine Analytical Processing) is the concept of analytical application in which the end user discovers problems and solutions by directly searching and organizing a database; the goal is to achieve quick data access and to provide a function for easy analysis.131
OLTP
OLTP (OnLine Transaction Process) is the processing mode in which messages are sent to the host computer from multiple terminals connected online to the host computer, which, according to the message received, in turn performs the process including access to a series of databases and returns the process results immediately to the terminals.
Databases used by OLTP are called business databases or, sometimes, operational databases.
These are terms in contrast with informational databases.
Below, we compare informational databases with operational databases:
Item for comparison
Target task
Data addition
Data updating
Processing mode
Main users
Business type
Period of data retention
Informational databases
Supporting decision-making
Yes
Generally no
OLAP
Management staff
Non-standard tasks132
Long term
Operational databases
General business tasks
Yes
Generally yes
OLTP
General workers
Standard tasks
Short term
Application Systems
Various systems that use databases are developed. Today, it is not an overstatement to say that most of the systems in operation use databases. Some of the great advantages for using databases are as follows:
• Data can be easily accumulated.
• Data can be managed in an integrated manner.
• Data can be easily processed.
• Data can be easily searched.
131
(FAQ) There have been exam questions concerning data warehouses. Know accurately the meanings of data warehouses, OLAP, and OLTP.
132
Standard/non-standard tasks: Standard tasks are those for which processing procedures are fixed, such as daily business procedures and daily input of sales data. Non-standard tasks are those for which processing procedures vary case by case. Creation of analysis documents, for instance, requires different processes depending on the purpose of use, so it is considered non-standard.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
122
2. Computer Systems
2.5.3 Multimedia Systems
¾
Points
¾
Examples of multimedia usage include AI, 3DCG, and pattern recognition. Multimedia application systems include Internet broadcasting and
VOD (Video On Demand).
Multimedia refers to handling not just characters and text but also mixtures of still images, moving pictures, audio, and other communication media. The term also refers to devices and software used in multimedia communication.
Artificial Intelligence (AI)
Artificial intelligence refers to a system that performs inference processes such as an expert system 133 and a machine translation system. A typical computer does no more than manipulating numerical values and performing logical operations. AI, on the other hand, performs inference processes centered on manipulating character strings. It is said that Lisp and
Prolog are languages suitable for developing software using AI technologies.
3-Dimensional Graphics (3D Computer Graphics, or
3DCG)
3D computer graphics134 involves creating virtual 3D space inside the computer, placing solid models in this space, and moving them around. It is embedded in movies, games, and animations. VRML135 is used to develop 3DCG.
The following table compares artificial reality and virtual reality.
Type
Artificial reality (AR)
Virtual reality (VR)
Explanations
It is the technology of creating a virtual world inside the computer, with a sense of reality.
Special equipment is not necessary to experience the artificial reality.
It is the technology of creating a fictitious world and having people experience and feel that world as though it were real.
3D vision using dedicated display units and special input equipment are used. Examples: pre-experience of virtual surgery, flight simulator, etc.
133
Expert system: It is a system created with the knowledge base of various specialists (experts) in a variety of fields; given certain conditions, the system applies the knowledge based on certain rules so that problems can be solved as if they were solved by the experts.
134
Computer graphics (CG): It is the technology of creating images via computers, or images made by such technology.
There are methods where the computer processes already existing images, and there are other methods where the computer creates images themselves. The latter method is called CGI (computer-generated images).
135
VRML (virtual reality modeling language): language specifications to describe 3DCG used on the Internet.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
123
2. Computer Systems
Multimedia Application Systems
There are many systems that apply multimedia. Now that WWW has brought in a graphical environment through the Internet, a variety of multimedia contents are available for use.
Internet broadcasting
This is broadcasting using multimedia on the Internet. With the use of streaming distribution technology, 136 programs are broadcasted in real-time on the Internet. Compared with conventional broadcasting business, equipment costs much less, and global information transmission is possible. In addition, on-demand service137 can also be provided, so we can tune in whenever we wish to watch a particular program.
Internet broadcasting comes in various formats. On-demand broadcasting stores the contents on a server and distributes them per request from a user. Live broadcasting (Internet live) distributes live programs, such as concerts, simultaneously to multiple users.
Non-linear editing
This is a method of video editing where images are digitized and video is produced in free order using a computer. It is easy to correct images, switch the order in which they appear, and create a different version. Incidentally, the conventional method in which video images are dubbed in the order of their completion is called linear editing.
Video on demand
This refers to the service of instantly sending a video program requested by the viewer via, for example, bidirectional CATV. The service provider stores many video programs on its video server and distributes the one requested by the viewer.
A video server can respond to simultaneous access by many viewers, and programs are requested to be sent from the beginning. Hence, it is necessary to construct an image database and connect it to individual mobile terminals and household receivers via broadband communication lines such as cable, wireless, etc.
136
Streaming: It is the technology of reading data and playing the data back immediately. It enables Internet broadcasting and playback of contents without waiting time. For streaming distribution, the line speed must exceed the amount of data; however, Internet lines are generally slow, so normally the data is compressed to enable real-time transmission.
Conventionally, playback used to be time-consuming since the data had to be downloaded first and then played back. With streaming, however, playback is done while data is being received.
137
On-demand: It is a function to provide what is requested whenever requested.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
124
2. Computer Systems
Quiz
Q1
What is an intranet?
Q2
What is a data warehouse?
A1
It is a network within a company applying the technologies of the Internet. Normally, a firewall is set up between the Internet and a company-wide network in order to prevent leakage of confidential information of the company.
A2
It is a company-wide database to support decision-making. It is the concept of storing a large amount of data in an organized manner to be used to help make business decisions. It is sometimes called an informational database.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
125
2. Computer Systems
Question 1
Q1.
There is a system which manages the file area in units of blocks. Each block contains eight sectors, and one sector is 500 bytes. How many sectors in total would be assigned to store two files, one consisting of 2,000 bytes and the other of 9,000 bytes?
Here, the sectors occupied by management information, such as directories, can be ignored. a) 22
b) 26
c) 28
d) 32
Answer 1
Correct Answer:
d
Files are saved in units of 8 sectors. Eight sectors, as shown below, are 4,000 bytes.
8 * 500 = 4,000 (bytes)
Hence, if one block is less than 4,000 bytes, all 4,000 bytes are used.
Next, we find the number of sectors necessary for each of the 2,000-byte file and 9,000-byte files. Capacity required for the 2,000-byte file
= 2,000 / 4,000=0.5 (block)
Æ 1 block (= 8 sectors) required
Capacity required for the 9,000-byte file
= 9, 000 / 4,000=2.25 (blocks)
Æ 3 blocks (= 24 sectors) required
Hence, to save files of 2,000 bytes and 9,000 bytes, the total number of sectors allocated to the two files is 32 as shown below.
Number of sectors required = 8 + 24 = 32 (sectors)
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
126
2. Computer Systems
Question 2
Q2.
Which of the following is the appropriate term for the process of breaking down data and storing it on multiple hard disks, as shown in the figure below? Here, b0 to b15 represent the sequence in which data is stored on the data disk in units of bits, and p0 to p3 represent the parity used to identify disk failure.
Control unit b0 b1
b2
b3
p0(b0 to b3)
b4
b5
b6
b7
p1(b4 to b7)
b8
b9
b10
b11
p2(b8 to b11)
b12
b13
b14
b15
p3(b12 to b15)
Data disk 1
Data disk 2
Data disk 3
Data disk 4
Parity disk
a) Striping
c) Blocking
b) Disk caching
d) Mirroring
Answer 2
Correct Answer:
a
Striping is to distribute one block of data onto two or more disks and write simultaneously. By striping, each block can be read and written in parallel, so the input/output speed increases.
Striping is defined as a technology for RAID. As shown in the figure above, this configuration contains a disk dedicated to the parity; such a configuration is called RAID2, RAID3, or
RAID4.
b) Disk cache is placed between a hard disk and the main memory; it is a buffer (buffer memory) to improve the apparent speed of the hard disk.
c) Blocking is to handle each logical set of multiple records as one physical record (block).
d) Mirroring is to prepare multiple disks and write the same data onto separate disks simultaneously, i.e., a multi-disk configuration. If one of the disks fails, the operation continues with the remaining disks only. This configuration is called RAID1.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
127
2. Computer Systems
Question 3
Q3.
Which of the following is arranged in the order of the effective memory access speeds from fastest to slowest?
With cache or w/o cache?
Cache memory
Access time
(ns)
-
Main memory
Hit rate
(%)
-
Access time (ns)
A
w/o cache
B
w/o cache
-
-
30
C
with cache
20
60
70
D
with cache
10
90
80
a) A, B, C, D
c) C, D, A, B
15
b) A, D, B, C
d) D, C, A, B
Answer 3
Correct Answer:
b
Cache memory (buffer memory) is memory which is placed between the CPU and the main memory to adjust speed differences between the two. The effective access speed can be increased by adding high-speed buffer memory, and by reading and writing on this cache memory as much as possible.
Let tc be the access speed of the cache memory, tm be the access speed of the main memory, and h be the hit ratio. The effective memory access speed is then calculated as follows:
Effective memory access speed = tc * h + tm * (1 – h)
The hit ratio is the probability that the data to be read is in the cache memory. The higher the hit ratio is, the faster the effective memory access speed becomes.
For A through D, we need to calculate the effective memory access time. For A and B, there is no cache memory, so the access time of the main memory is the effective memory access time.
Here, we can consider h = 0. The numbers below indicate the order of each, from fastest to slowest (from the shortest effective memory access time to the longest).
A: 15 (ns)
(1)
B: 30 (ns)
(3)
C: 0.6 * 20 + (1 – 0.6) * 70 = 40 (ns) (4)
D: 0.9 * 10 + (1 – 0.9) * 80 = 17 (ns) (2)
Hence, if we arrange Memory A through D from fastest to slowest in terms of the effective memory access time, the order is “A, D, B, C.”
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
128
2. Computer Systems
Question 4
Q4.
When a certain file was copied from one directory to another on a hard disk in a PC, file fragmentation occurred. Which of the following is an appropriate description concerning this situation?
a) The fragmentation can be eliminated by performing a physical dump of the entire disk and then restoring the disk.
b) Access time will be longer even for some files other than the file in which the fragmentation occurred.
c) If the file in which the fragmentation occurred is copied again, the fragmentation in the copy destination may get worse, but it will never be eliminated.
d) Even if fragmentation has occurred, the size of the file is still the same as that of the original one.
Answer 4
Correct Answer:
d
Fragmentation means that this file saved on the hard disk could not secure one continuous area and thus is saved across multiple blocks. When fragmentation occurs, various parts of the hard disk must be accessed, reducing the processing efficiency. However, the file size does not change, as the file is simply saved in a divided manner.
a) If the disk is physically copied, the situation does not change. It must be copied logically.
b) One file was copied on the hard disk on which files had already been saved, so no other file except the one that was copied was affected. Physically nothing was changed.
c) The fragmentation will be solved if the copy destination has a continuous empty area whose size is larger than the file size.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
129
2. Computer Systems
Question 5
Q5.
The table shown below gives processing times for a CPU and I/O devices to execute 5 stand-alone tasks. Which task can be executed simultaneously with the “High” priority task so that the CPU idle time from task execution start to end can be zero? Here, each task uses a different I/O device and is performed concurrently. The overhead of the OS can be ignored.
Unit: ms
Priority
Stand-alone task processing time
High
CPU (3) →I/O (3) → CPU (3) → I/O (3) → CPU (2)
a)
Low
CPU (2) →I/O (5) → CPU (2) → I/O (2) → CPU (3)
b)
Low
CPU (3) →I/O (2) → CPU (2) → I/O (3) → CPU (2)
c)
Low
CPU (3) →I/O (2) → CPU (3) → I/O (1) → CPU (4)
d)
Low
CPU (3) →I/O (4) → CPU (2) → I/O (5) → CPU (2)
Answer 5
Correct Answer:
c
We check the operation of each “low-priority” task to see which one uses the CPU while the
“high-priority” task is not using the CPU. The input/output units are different, so there is no waiting for the input/output (I/O) units.
The “high-priority” task uses I/O units twice, each for 3 ms. If the use of the CPU by the
“low-priority” task takes exactly 3 ms, the CPU has no idle time. With this said, let us now consider each task listed in the answer group.
For instance, consider Task (a). During the 3-ms period when the “high-priority” task uses an
I/O unit for the first time, the “low-priority” task uses the CPU for 2 ms. Hence, 1 ms of CPU idle time will result.
Similarly, any “low-priority” tasks like (b) and (d), with 2 ms of CPU use, will cause CPU idle time. In contrast, Task (c) has 4 ms of CPU use, but this comes in the end, so no idle time will occur if this is after the completion of the “high-priority” task.
Hence, the “low-priority” task that can completely eliminate the CPU idle time until the execution of both tasks is completed is “CPU (3), I/O (2), CPU (3), I/O (1), CPU (4).”
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
130
2. Computer Systems
Question 6
Q6.
The state transition diagram below shows a task (process) state transition on a multitasking computer. When does the task state change from the running state to the ready state?
Running state
Ready state
a)
b)
c)
d)
Waiting state
A task with higher priority compared to its own has moved to the ready state.
A task has been generated by the job scheduler.
An I/O operation has completed.
An I/O operation has been requested.
Answer 6
Correct Answer:
a
When a task (process) is generated, it proceeds to the ready state and is entered into the queue.
After that, it goes into the running state depending on the task priority and goes into the waiting state when an I/O operation occurs. Such changes in the task state are called the state transitions of the task.
(5)
(1)
(3)
task generated Running state B
A
Task eliminated C
E
Ready state F
D
Waiting state (4)
(2)
We now explain the state transitions of tasks. The numbers in parentheses ( ) are for state explanations, and A through D are for transition explanations.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
131
2. Computer Systems
No.
(1)
(2)
(3)
(4)
(5)
State
Task generated
Ready
Running
Waiting
Task eliminated
Transition
A
B
C
D
E
F
Explanation
A task generated is registered into the task queue.
Waiting to be authorized to use the CPU
Authorized to use the CPU and is currently being executed
Waiting for completion of an event such as an I/O operation
A task is completed and no longer necessary; it is removed from the task queue.
Explanation
A task is generated; moved to the ready state
By priority, moved to the running state
Moved to the waiting state (waiting for I/O, etc.) due to an event
Moved to the ready state after completion of an event
Moved to the ready state due to another task with high priority
The task in the running state is completed.
As you can see from these tables, transition from the running state to the ready state is transition E. This is when a task whose priority is higher than the task being executed goes into the ready state.
b) This describes transition A in the state transition figure for the tasks.
c) An I/O operation is an operation of input or output. When this I/O operation is completed, an
I/O interruption occurs and the task moves into the ready state. This describes transition D in the state transition figure for the tasks.
d) An I/O operation is requested when the task issues an instruction for input or output. If this happens, a supervisor call occurs, and the task moves from the running state to the waiting state to wait for the completion of the I/O. This is transition C in the state transition figure for the tasks.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
132
2. Computer Systems
Question 7
Q7.
Which of the following is an appropriate statement concerning a client/server system?
a) The client and the server must use the same kind of OS.
b) The server sends data processing requests and the client processes those requests.
c) A server can support a client function that enables it to request processing of another server if necessary.
d) The server functions must be allocated to different computers, such as a file server and print server.
Answer 7
Correct Answer:
c
A client server system is a system made up of processing units called clients and servers. A client performs data input/output and other processes through a server while a server controls all input and output that depend on the hardware according to its type. Normally, a client unit is a unit equipped with data processing functions such as a personal computer or a workstation, so it can perform applications on its own. In fact, it also performs processes only a client can perform, such as displaying text and drawing figures. A server, on the other hand, accesses databases and performs printing processes in response to requests by clients. Further, if a server is not able to perform a process, it can request another server to do that. Here, the first server becomes a client because it is requesting another to perform a process.
a) Different operating systems do not cause any problems as long as the protocol is established.
We can have a combination of servers with UNIX and clients with Windows.
b) It is a client that sends requests for processing. It is a server that performs the requested processes. d) In a small-scale system, a server and a client can even be the same. X Windows of UNIX is an example of this type. If the clients and servers are built on the same platform (OS) and can be connected via a network, there is no inhibitory effect.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
133
2. Computer Systems
Question 8
Q8.
When comparing a distributed processing system, which consists of multiple computer systems located in a wide area, with centralized processing systems that operate in a single center, which of the following is the most appropriate feature of centralized processing systems?
a) In the event of a disaster or a failure, recovery work can be conducted in a centralized fashion in the center, avoiding the risk of a long shutdown of the entire system. b) Since the system is collectively managed, it is easy to satisfy requests for additions or changes to system functions, and the accumulation of backlog seldom occurs.
c) Data consistency is easily maintained and managed through the centralized implementation of measures in the center.
d) Although the operation and management of hardware and software resources become complex, expansion taking advantage of new technologies is easy.
Answer 8
Correct Answer:
c
A centralized processing system is the idea of centralizing processes to a general-purpose large computer (host computer). Since the equipment is centralized at one location, the management is easy. However, if the host computer fails, it is possible that the system gets shut down for a long period of time.
a) Since the host computer performs all processing, the system gets shut down until the host computer is recovered. If this is a serious failure, it is possible that the downtime of the system becomes long.
b) In a centralized processing system, all requests must be answered by the host computer alone.
If the contents of the requested items vary significantly in level, the host computer cannot meet all those requests, causing backlog accumulation. Incidentally, backlog can also refer to a system waiting to be developed.
d) A centralized processing system processes all tasks at the host computer, so it is cumbersome to respond to each type of business task separately. Even if we want to extend the system, often there are tasks that cannot be suspended. With the introduction of a new technology, some tasks may not be able to be processed any longer.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
134
2. Computer Systems
Question 9
Q9.
For one job, which of the following formulas appropriately expresses the relationship between turnaround time, CPU time, I/O time, and process waiting time? Here, other types of overhead time are ignored.
a)
b)
c)
d)
Process waiting time = CPU time + Turnaround time + I/O time
Process waiting time = CPU time - Turnaround time + I/O time
Process waiting time = Turnaround time - CPU time - I/O time
Process waiting time = I/O time - CPU time - Turnaround time
Answer 9
Correct Answer:
c
Process waiting time is the time until the start of a CPU process or an I/O process of the job entered into the computer system. Turn-around time (TAT) is the time interval from submitting a job to receipt of the results. This concept is mainly used in batch processing. TAT includes both the CPU processing time and I/O time. Hence, process waiting time is obtained by subtracting CPU processing time and I/O processing time from TAT.
In the figure below, the shaded areas represent process waiting time.
Job entered
Job completed
CPU
processing
I/O processing CPU processing I/O processing Turn-around time
Hence, for one job, the formula expressing the relationship among turn-around time, CPU time,
I/O time, and process waiting time is as follows:
Process waiting time = Turn-around time – CPU time – I/O time.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
135
2. Computer Systems
Question 10
Q10. LAN facilities are installed as shown in the figure below. Using the server connected to LAN3, the client on LAN1 is performing a business application. Data transmission is normally performed via Router 1. If a failure occurs in Router 1, Routers 2 and 3 are used for transmission between LAN 1 and LAN 3. What is the availability of the
LAN equipment connecting LAN1 and LAN3? Here, the failure rate of each router is
0.1, no switch-over time is required in case of a failure, and failures in LAN facilities other than the routers are not taken into account.
Server
LAN 3
Router 3
LAN 2
Router 1
Router 2
LAN 1
Client
a) 0.729
b) 0.981
c) 0.990
d) 1.000
Answer 10
Correct Answer:
b
Note that “Router 1” and “the serial connection of Routers 2 and 3” are in parallel connection.
Since the failure rate of each router is 0.1, the availability is 0.9. Hence, the configuration is as shown below. The values in the boxes are availability of each unit.
Configuration 1
Router 2
0.9
Router 3
0.9
Client
Server
Router 1
0.9
Availability of Configuration 1 = 0.9 * 0.9 = 0.81
Availability of the entire system = 1 – (1 – 0.81) * (1 – 0.9) = 0.981.
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
136
2. Computer Systems
Question 11
Q11. Which of the following is an appropriate statement concerning VR (Virtual Reality)?
a) Using technology such as CG, VR expresses the world created inside a computer as if it were the real world.
b) For the purpose of improving GUI, it does not display an image incrementally from the top, but first displays a rough mosaic-like image and gradually sharpens it.
c) VR tests whether or not hypothetical results can be obtained from computer simulations such as those of wind tunnel tests used for automobile or aircraft design. d) VR makes abilities such as human recognition and inference possible on a computer. Answer 11
Correct Answer:
a
Virtual reality (VR) is to artificially create a sense of reality by combining CG (Computer
Graphics) and sound effects. To appeal to the senses, we may also use dedicated display units such as a head-mounted display (HMD) for 3D vision and special input units such as data gloves. In addition, the response to the images is returned to the user, further enhancing the sense of reality.
b) This describes interlace.
c) This describes simulation.
d) This describes AI (Artificial Intelligence).
FE Exam Preparation Book Vol.1
-- Part1. Preparation For Morning Exam --
137
You May Also Find These Documents Helpful
-
On the basis of the following data for Seller Co. for 2008 and the preceding year ended December 31, 2007, prepare a statement of cash flows. Use the indirect method of reporting cash flows from operating activities. Assume that equipment costing $125,000 was purchased for cash and equipment costing $85,000 with accumulated depreciation of $65,000 was sold for $15,000; that the stock was issued for cash; and that the only entries in the retained earnings account were net income of $51,000 and cash dividends declared of $13,000.…
- 618 Words
- 6 Pages
Satisfactory Essays -
Microsoft. (2012, April 26). Window Server Creating A Site Link Design. Retrieved from Tech Net.Microsoft.com: http://technet.microsoft.com/en-us/library/cc732837(v=ws.10).aspx…
- 670 Words
- 2 Pages
Satisfactory Essays -
Ga atomic mass is 69.72g/mol. Let x be the % of 69Ga as a decimal and therefore 1.000 – x is the % of 71Ga as a decimal.…
- 562 Words
- 3 Pages
Satisfactory Essays -
Together with AK/ADMS 4562 3.0, introduces students to the principles and practice of Canadian taxation and related tax planning, to provide a basic understanding of the Canadian Income Tax Act and its GST implications in relation to individuals.…
- 4726 Words
- 19 Pages
Good Essays -
Employee A has asked to take a leave of absence from work for a family medical reason. Employee A was granted the time off to be with his spouse for the premature birth and care of their twins. After 11 weeks off from work, he has put in a request to return to his position with pay for the time he had off.…
- 748 Words
- 3 Pages
Good Essays -
This exam will be a combination of multiple choice & matching questions and a few short essay responses. For best results in preparing for the upcoming exam, focus your studies on the following historical items, events and individuals. Your textbook may help with some of the things listed, but this exam is more so drawn from class lectures and power point presentations.…
- 510 Words
- 3 Pages
Satisfactory Essays -
Problem Statement: A football league scores points for field goals as five points and touchdowns as three points. They want to know what the highest amount of points that is impossible to score.…
- 591 Words
- 3 Pages
Good Essays -
Beginning at the end of June of 1977, John, Yoko, and son Sean will spend four months visiting ________, including local tourist attractions and Yoko 's family.…
- 1254 Words
- 9 Pages
Satisfactory Essays -
Identify legislation relating to health and safety in a health or social care work setting…
- 364 Words
- 5 Pages
Satisfactory Essays -
two major tribes were the Ostrogoth’s and the Lombard’s where power shifted several times, And the…
- 1759 Words
- 8 Pages
Powerful Essays -
1. In the paternalistic model of doctor-patient relationship, power resident exclusively with Doctors; in the engineering model, it resides entirely with Patients.…
- 892 Words
- 4 Pages
Satisfactory Essays -
As the next generation of colonists moved westward to find new, fertile land, they encountered plentiful acreage at cheap prices. Frontier families lived with the bare necessities acquired through subsistence farming, created a widely dispersed society of equals, and were subjected to a disorganized existence without organized law and order, community institutions, or organized churches. Thus, frontier communities became volatile and violent places where deep divisions festered between its residents and those of the eastern seaboard.…
- 2006 Words
- 8 Pages
Good Essays -
3. Be able to plan shared approaches to the care and support of individuals with families…
- 714 Words
- 4 Pages
Good Essays -
The foundation of orthodoxy and the New Testament Canon are connected to one another because they were used and based on scriptures and testimonies written by the Apostles. Christians believed the writings of the Apostles because in their minds there was no reason to question those authoritative writings since they had documented the time that they spent with Jesus and the instructions that he gave to each Apostle as well as giving them instructions on the church. Bruce Metzger, states that “The issues were whether the Rule of Faith determined the extent of the canon or was shaped by it and what constituted apostolicity and authority of Scripture.”(Metzger,1987:127). According to Metzger, “The New Testament book opens with a judicious and critical analysis of scholarly writings on the subject of the canon prior to and during the twentieth century.” (Metzger, 1987:127). The early church was dealing with many challenges internally and externally. There were movements that were within the church that questioned the orthodoxy and the New Testament canon. One of the movements included the Gnosticism, with its claim to esoteric knowledge that supplements and basically alters the outlook of the writers of the New Testament: the attempt by Marcion to sever the connection with Judaism by discarding the Old Testament and reducing the authoritative Christian writings to an expurgated edition of Luke and the letters of Paul; and Montanism, which claimed to supplement the New Testament by revelatory insights and discourses (Metzger, 1987:127). A major factor in the East was the canonical letter of Athanasius of Alexandria in 367, in which…
- 1837 Words
- 6 Pages
Powerful Essays -
_____ Psychology and "common sense" lead to the same conclusions about behavior and mental processes.…
- 6504 Words
- 60 Pages
Good Essays