Memory Management - Essay Example

 Memory Management. 1. Compare between fixed partition memory organization with single queue and with multiple queues in terms of the basic idea, advantages, and disadvantages. Please use table format. Memory organization needs to be done in order to allocate proper memory slots to the ever emerging processes. The memory of a computing system is required to be organized such that each process gets the best available slot in accordance with its needs and also with respect to the available memory. The partitioning that is done in memory for its allocation can be fixed. There are two possible methods whereby which queues may be used to bring processes into the memory for execution. The first method is by using a single queue and the second method I by using multiple queues as shown in the fig. 1. Figure 1 Fixed memory Partitioning. Using multiple queues Using Single Queue Mechanism The first available memory location that is large enough to hold the process is allocated All processes get lined up into a single queue. The partition closest to the process that can hold it is assigned to that process. Advantages Load is balanced. Arrived processes are transferred evenly into all memory partitions. As soon as a memory partition gets free it is assigned to a waiting process. The allocation process is very quick. The allocation process does not waste memory space. Disadvantages Any memory size that may be more than the size of the process gets lost. Moreover the queues of small partitions are often lined up with several processes while those with larger partitions are often left untouched. The drawback is that even though large sizes of memory are free small processes are not allocated that memory and they stay lined up in the queues of smaller memory sized partitions waiting for their turn to come. A large partition may be wasted on a small job resulting in loss of large amounts of memory. Unbalanced load. Some queues may be empty while others are fully loaded. 2. What is the swapping technique? Why is it needed for multiprogramming environment? Swapping generally means interchanging the contents of two different spaces using a third temporary empty content space. Swapping is used as a memory management technique for multiprogramming environment. For a program to execute it is mandatory to acquire physical memory or main/primary memory. In a uni-processor environment multiple programs can simultaneously be executed by overlapping an I/O of a program with the execution of another program. Due to the limited size of physical memory it is very difficult to accommodate all of the required programs in main memory. To accommodate new programs currently running programs are swapped out on the basis of their activity status. For example when a program asks for data from secondary storage, it has to wait for complete input. Meanwhile due to Direct Memory Access the CPU remains available for other programs. The operating system thus swaps out the waiting program with the one that needs to be executed with the CPU. Upon completion of data transfer the earlier program is invoked and swapped into the memory again. This is how swapping works for multiprogramming environment. 3. Discuss the main function(s) of each of the following operating system parts: a) Memory management Unit The Memory Management Unit is said to serve as a buffering agent between the Central Processing Unit and the memory of the system. The functions that are performed by this unit may roughly be categorized into three categories. These are: 1. Operating System Memory Management, 2. Hardware Memory Management and 3. Application Memory Management. Specifically speaking, it translates address from their virtual address format into their physical address format. b) Linker “Performs the function of linking the libraries to the object code in order to transform the object code into an executable machine.” (cool interview, 2012). Simply put, it combines all the related attachments of a user process such as the processes that run with it, its libraries etc into a single package so that memory is allocated to them altogether. c) Page replacement algorithm The page replacement algorithms are integral with their functionality. They assist primarily when a page fault occurs. They perform the function of deciding which page can be removed from the memory so that space can be made for a page which is required for execution by the processor. The page replacement algorithm ensures that the page being removed is not currently in the use of the CPU. d) Loader The function of the loader is to look for a process that is not currently in memory but is needed by the CPU. It allocates memory space to the said process and thus brings it into memory so that the CPU can utilize it. The execution is then given to the process that is loaded by the loader. 4. List the advantages and disadvantages of Fixed partition and Variable partition memory organizations. Fixed Partition Memory Organization Variable Partition Memory Organization Advantages Saves the hassle of runtime memory partitioning. Facilitates quick memory allocation. Simple to implement. Robust. Offers three authentic memory allocation algorithms namely Worst Fit, First Fit and Best Fit. Facilitates random program sizes by making available required sized memory partitions at runtime. Sometimes accommodates more programs into memory in an optimized manner. Comparatively lesser internal fragmentation. Allows compaction as a remedy against external fragmentation. Disadvantages Allows contiguous allocation of memory only. Not optimized. High external and internal fragmentations. Runtime adjustments for optimization are not allowed. Sometimes exhibit maximum fragmentation, due to which the execution slows down extremely. Allows contiguous memory allocation only. Runtime partition size determination and creation affects the time efficiency of operating system. Exhibits external fragmentation. Sometimes worst as compared to fixed partition allocation due to dispersed non-contiguous empty left over spaces. Compaction can only be applied to runtime bounded processes/programs. 5. When is page fault? Describe the actions taken by the operating system when it occurs? Demand paging is an advanced memory allocation mechanism. It is an extension of a non-contiguous memory allocation mechanism that is simply known as Paging. In paging the physical memory is segmented into multiple fixed sized chunks that are referred to as frames. Similarly the logical memory or program content itself is also divided into same sized chunks to fit into frames, hence referred to as pages. The size of frames and pages is determined in power of two. The pages are then placed into different available frames randomly and the record of storage is kept as page table. Page table contains the page number along with its respective frame number into which that page resides. It is impossible to accommodate all the pages of all programs/logical memories into the main memory. A modern technique introduces an illusion of extension of physical memory using secondary storage. This technique is named as Virtual Memory. A specially designated area on secondary storage is fragmented as frames as a virtual extension of physical memory. All programs are brought into this secondary storage area to maintain the illusion that the programs have arrived into physical memory. Demand paging means paging on demand. Virtual memory and demand paging go hand in hand. Whenever a demand for a page arises through an initiation of respective logical memory address, the operating system searches for it into the page table for main memory. Page Fault occurs when a page is not found in main memory. On page fault the operating system invokes page fault service routine which performs the following steps. 1. Searches for the respective page into virtual memory. 2. If found, then loads it into one of the physical memory frames. 3. Updates the respective page table. 4. Hands over the control to operating system that again reinitiates the request for the same page. This time the page is found in memory and gets executed. The above mentioned steps cause a respective delay that is referred to as page fault service delay. 6. Is it possible that a segment could belong to the address space of two different? Processes? Why? Yes, it is possible that a single segment may belong to the address space of two different processes. This is because segment tables are a collection of base-limit registers. When two or more entries of two or more jobs in their respective segment tables point to the same memory location that memory segment gets shared. The two segment tables having aiming to access the same memory location must have the same base pointers. Also, it is imperative that the segment number that is being accessed must be the same in both the processes. 7. List the advantages and disadvantages of one dimensional and two dimensional memory organizations. One Dimensional Memory Organization Two Dimensional Memory Organization Advantages All the information associated with a process is allocated only one memory partition. The memory does not have to jump to separate locations to access the sister tasks of a single process and they can all be accessed by accessing just one memory partition. This reduces the searching execution for them. Independent address space units called segments are assigned to all sister processes of a job. Segments are independent of their sizes and may grow separately without affecting other segments. Thus one segment growth does not affect the growth of another process’s segments. Linking up of independently compiled procedures becomes convenient in the two dimensional memory organizations. Data stored in a single segment may be accessed by other segments as well. Disadvantages A process size may grow enormously and may bump into other jobs that may similarly be growing into sizes as well. Large sized segments form sometimes. These segments may not be kept in memory any longer due to their increased size. Two dimensional memory organization results in holes along with segments. These holes need to compacted further more so as to make efficient use of memory. WORKS CITED: Cool Interview. 2012. Web. March 13, 2012. < http://www.coolinterview.com/interview/2616/> Tanenbaum, Andrew. Modern Operating Systems. Upper Saddle River (N.J.: Pearson Prentice Hall, 2008. Print. Silberschatz, Abraham, Peter B. Galvin, and Greg Gagne.Operating System Concepts Essentials. Hoboken, NJ: John Wiley & Sons Inc, 2011. Print Read More