Memory Management

Intro to Operating Systems
POS
Requirements of Memory Management
Memory Management is the most highly used resource on your computer; different operating systems all have a different approach on how to use memory and common factors when using memory. Every process needs it to process a code or function. More so memory management allow the operating system to reserve parts of memory to programs when they are requested and complete as many processes into memory. Memory Management has gotten more complicated throughout the years with advancement of operating systems and hardware to be able to provide faster computing of processes and applications that demand a lot of memory along with speed and also the improvement of just how the memory is allocated. Operating Systems utilized the ability use memory located on RAM chips, cache chips, and including the hard drive with the use of page file or sometimes called Swap files. Single Partition mono programming was an earlier example of memory management utilized by MS-DOS where only a single program would be running and keep its process in memory, and the other portion of memory reserved for the operating system and device drivers. Of course with advancement of faster hardware and the eventual progression of technology having just one program running at a time was efficient or fast even though it was stable so another method of management was adopted. Multiprogramming is the process run multiple programs simultaneously, and swaps between processes as the user chooses. The Major reason programs would be loaded to memory when they were in use for the simple fact memory is faster and located closer to the CPU and performed dramatically faster than the hard drive trying to perform the process of memory management. Most processes all go according to time-blocks where every process waits in memory for their turn to occur to process their request or process, more programs in line in memory the more CPU utilization is occurring. There are different ways to position code and address memory such as absolute code where addresses are hard coded into the program then there is position dependent code where they are addressed to their relative location. Another method is dynamically relocate-table code where the use of a compiler to represent locations that offset locations where absolute memory exists. Another different approach of memory management is logical addressing which will translate any address reference from logical to physical address, and the use of a memory management unit performs this real-time. Another key to know about processes is the fact they always grow so over time process slowly get bigger and bigger so its always a good practice to allocate extra memory for all processes, if a processes become to big only options are to swap to a bigger hole or swap it on the disk or kill the process taking up the memory slowing down the CPU looking for additional space to grow. Of course with growing process, segmentation is taken to account in a variation of variable partition multiprogramming by breaking up the process into smaller chunks so at least the holes it can fill can be smaller and found throughout the system to fulfill its requirements. The memory management unit has to handle the segments by handling the segment registers associated with code, data, stack, and heaps which are still used in today’s 64-bit systems. Allocation algorithms is how memory management chooses spots to fit the segment for processes and code, different approaches such as going with first come first serve or what is the next one or even the best fit or just a quick fit but drawback is find other places to merge can be difficult. Page based virtual memory is a technique which allows the process to feel like it has entire address space on the processor and allows memory allocation to be non contiguous and has no external fragment problems and also lets us keep part of process were using in memory and the rest on the disk. On top of managing logical and physical addresses, memory management also provides protection and statistics, which include valid/invalid if there is a frame mapped to a location. Read-only if the process is allowed to write on the page, execute to process a code on a page including dirty, and accessed page file locations. With page virtual memory process can get available discontinuous pages as they need additional memory. The table for a process can give the illusion of continuity by making the virtual address space continuous, there are processes that bypass memory management such as DMA (direct memory access) which directly services devices over the CPU and MMU to provide service when the CPU is being utilized for faster service to handle requests and is built into the operating system also to provide this function.
Memory management has lots of features and all different operating systems take approaches on what is best for their code to utilize memory processes. With any computer system there is main memory and secondary memory, such physical memory that is permanent to a computer is RAM which comes in various types such as regular RAM (random access memory), SRAM (static random access memory), DRAM (dynamic access memory ram) and includes a memory controller to operate each type of memory. Of course main memory is faster and temporary storage and secondary memory is slower but more reliable. Another type of physical memory being utilized more and more is cache memory, which is located close to the CPU and allows the CPU to execute processes faster, which is utilized by the operating systems to perform functions with a high priority. Memory management usually either optimizes look up time better or how to better use space to allocate memory addresses. Understanding memory management is key when administering different operating systems with different methods on how they allocate memory and knowing common issues that might arrive depending on the algorithm they have to utilize since every system has its pro’s and con’s and can better help when choosing an operating system depending on the type of application purposes it will be used for.

WORK CITED
Wilson, P.R.; Johnstone, M.S.; Neely, M.; Boles, D. (1995).”Dynamic Storage Allocation: A Survey and Critical Review” Memory Management: International Workshop, Iwmm '95, Kinross, Uk, September 27–29, 1995: Proceedings (Springer). ISBN 978-3-540-60368. Retrieved 2008-01-06.
Berger, E.D.; Zorn, B.G.; McKinley, K.S. (2002). “Reconsidering custom memory allocation”. Proceedings of the 17th ACM SIGPLAN conference on Object-oriented programming, systems, languages, and applications. ACM Press New York, NY, USA. pp. 1–12.

Cited: Wilson, P.R.; Johnstone, M.S.; Neely, M.; Boles, D. (1995).”Dynamic Storage Allocation: A Survey and Critical Review” Memory Management: International Workshop, Iwmm '95, Kinross, Uk, September 27–29, 1995: Proceedings (Springer). ISBN 978-3-540-60368. Retrieved 2008-01-06. Berger, E.D.; Zorn, B.G.; McKinley, K.S. (2002). “Reconsidering custom memory allocation”. Proceedings of the 17th ACM SIGPLAN conference on Object-oriented programming, systems, languages, and applications. ACM Press New York, NY, USA. pp. 1–12.