221601304_FOS_UNIT_4.pdf

0301304 FUNDAMENTAL OF OPERATIONG SYSTEM UNIT MODULES WEIGHTAGE 1 INTRODUCATION TO OPERATING SYSTEM 20 % 2 PROCESS MANAGEMENT 20 % PROCESS COMMUNICATION AND 3 20 % S...

0301304 FUNDAMENTAL OF OPERATIONG SYSTEM UNIT MODULES WEIGHTAGE 1 INTRODUCATION TO OPERATING SYSTEM 20 % 2 PROCESS MANAGEMENT 20 % PROCESS COMMUNICATION AND 3 20 % SYNCHRONIZATION 4 MEMORY MANAGEMENT 20 % FILE MANAGEMENT , DISK MANAGEMENT , 5 20 % SECURITY AND PROTECTION UNIT – 4 Memory Management Basic Memory Management – Introduction – Basic Concepts Static and Dynamic Allocation Logical and Physical Addresses Fixed and Variable Memory Partitioning Fragmentation Swapping – Contiguous Memory Allocation Compaction Memory Allocation Techniques – First fit – Best fit – Worst fit UNIT - 4 Memory Management 2 UNIT – 4 Memory Management Paging Concept Segmentation Virtual Memory – Introduction – Need for virtual Memory – Demand Paging – Page Replacement Algorithm FIFO LRU UNIT - 4 Memory Management 3 UNIT – 4 Basic Mamory Management The multi programming concept of an OS gives rise to another issue known as memory management. Memory, as a resource, needs to be partitioned and allocated to the ready processes, such that both processor and memory can be utilized efficiently. The division of memory for processes needs proper management, including its efficient allocation and protection. There are two types of memory management : – Real memory (Main Memory) – Secondary memory UNIT - 4 Memory Management 4 UNIT – 4 Basic Mamory Management Memory allocation is generally performed through two methods: – Static Allocation – Dynamic Allocation Static Allocation – The allocation is done before the execution of a process. Dynamic Allocation – If memory allocation is deferred (at later time) till the process starts executing, it is known as Dynamic Allocation. UNIT - 4 Memory Management 5 UNIT – 4 Basic Mamory Management Static Allocation – There are two instances when this type of allocation is performed: When the location of the process in the memory is known at compile time, the compiler generates an absolute code for the process. When the location of the process in the memory is NOT known at compile time, the compiler does not produce an actual memory address but generate a relocatable code (Relocatable code is software whose execution address can be changed), that is, the addresses that are relative to some known point. UNIT - 4 Memory Management 6 UNIT – 4 Basic Mamory Management Dynamic Memory Allocation – In Multi-Programming, Modern OS adopt dynamic memory allocation method. – In this method, two types of addresses are generated. Logical Addresses Physical Addresses UNIT - 4 Memory Management 7 UNIT – 4 Basic Mamory Management Logical Addresses – In dynamic allocation, the place of allocation of the process is not known at the compile time and load time. – The processor, at compile time, generate some address, known as logical addresses. – The set of all logical addresses generated by the compilation of the process is known as logical address space. UNIT - 4 Memory Management 8 UNIT – 4 Basic Mamory Management Physical Addresses – Logical addresses need to be converted into absolute addresses at the time of execution of the process. – The absolute addresses are known as physical addresses. – The set of physical addresses generated, corresponding to the logical addresses during process execution, is known as physical address space. – When a process is compiled, the CPU generates a logical address, which is then converted into a physical address by the memory management component to map it to the physical memory. UNIT - 4 Memory Management 9 UNIT – 4 Basic Mamory Management Swapping – There are some instance in multi programming when there is no memory for executing a new process. – In this case, if a process is taken out of memoy, there will be space for a new process. UNIT - 4 Memory Management 10 UNIT – 4 Basic Mamory Management Swapping – It raise some question : Where will this process reside? Which process will be taken out? Where in the memory will process be brought back? UNIT - 4 Memory Management 11 UNIT – 4 Basic Mamory Management Swapping – It raise some question : Where will this process reside? – Secondary storage (generally Hard disk) known as backing store. – The action of taking out a process from memory is called swap-out. The process is known as a swapped-out process. – The action of bringing back the swapped-out process into memory is called swap-in. UNIT - 4 Memory Management 12 UNIT – 4 Basic Mamory Management Swapping – A separate space in the hard disk as swap space, is reserved for swapped out processes. UNIT - 4 Memory Management 13 UNIT – 4 Basic Mamory Management Swapping – It raise some question : Which process will be taken out? – In round robin process-scheduling, the processes are executed, according to the their time quantum. If the time quantum expires and a process has not finished its execution, it can be swapped – out. – In priority – driven scheduling, if a higer – priority process wishes to execute, lower – priority process in memory will be swapped out. – The blocked processes, which are waiting for an I/O, can be Swapped out. UNIT - 4 Memory Management 14 UNIT – 4 Basic Mamory Management Swapping – It raise some question : Where in the memory will process be brought back? – There are two options to swap The first option is to swap – in the process at the same location, if there is compile time or load time binding. Other option is to place the swapped -in process any where there is space. Need to relocation. UNIT - 4 Memory Management 15 UNIT – 4 Basic Mamory Management Swapping Time – A time take to acces the hard disk. Example : – A process of size 200 MB needs to be swapped into the hard disk. But there is no space in memory. A process of size 250 MB is lying idle in memory and therefore, it can be swapped out. How much swap time is required to swap-in and swap-out the processes if: Average latency time of hard disk = 10 ms Transfer rate of hard disk = 60MB / s UNIT - 4 Memory Management 16 UNIT – 4 Basic Mamory Management Solution : – The transfer time of the process to be swapped-in to hard disk = 200 / 60 = 3.34 s = 3340 ms – The swap time of 200 MB process = 3340 + 10 = 3350 ms – The transfer time of the process to be swapped-out form memory = 250 / 60 = 4.17 s = 4170 ms – The swap time of 250 MB process = 4170 + 10 = 4180 ms – Total swap time = 3350 + 4180 = 7530 ms UNIT - 4 Memory Management 17 UNIT – 4 Basic Mamory Management Fixed and Variable Memory Partitioning – Fixed Partitioning In this method of partitioning, the memory is partitioned at the time of system generation. – Variable Partitioning In this method, partitioning is not performed at the system generation time. The partition are created at runtime, by the OS UNIT - 4 Memory Management 18 UNIT – 4 Basic Mamory Management Fragmentation – Internal Fragmentation When a process is allocated to partition, it may be possible that its size is less than the size of partition. It leave a space after allocation, which is unusable by any other process, this wastage of memory, internal to a partition is known as internal Fragmentation. UNIT - 4 Memory Management 19 UNIT – 4 Basic Mamory Management UNIT - 4 Memory Management 20 UNIT – 4 Basic Mamory Management Fragmentation – External Fragmentation When allocating and de-allocating memory to the processes in partitions through various method. It may possible that there are small spaces left in various partitions throughout the memory. This memory space is known as External Fragmentation. UNIT - 4 Memory Management 21 External Fragmentation UNIT - 4 Memory Management 22 UNIT - 4 Memory Management 23 UNIT – 4 Continuous Memory Allocation In older systems, memory allocation is done by allocating a single contiguos area in memory to the processes. But in multi -programming system, memory was divided into two partitions. – One for the Os – Other for the User process UNIT - 4 Memory Management 24 UNIT – 4 Continuous Memory Allocation In Multi-user systems, more processes are accommodated by having multiple partitions in the memory. UNIT - 4 Memory Management 25 UNIT – 4 Contiguous Memory Allocation Here process is allocated a contiuous memory in a single partition. Thus the memory partition, which fits the process, is searched and allocated. The memory partition which is free to allocate, is known as a hole. When the process terminates, the occupied memory becomes free and the hole is available again. As soon as a process terminates, a hole becomes free, and is allocated to a waiting process. UNIT - 4 Memory Management 26 UNIT – 4 Compaction Compaction help to control memory wastage, occurring in dynamic partitioning. The OS observes the number of holes in the memory and compacts them after a period, so that a contiguous memory can be allocated for a new process. The compaction is done by shuffling the memory contents, such that all occuupied memory region is moved in one direction, and all unoccupied memory region in the other direction. This results in contiguous free holes, as a single large hole. UNIT - 4 Memory Management 27 UNIT – 4 Compaction UNIT - 4 Memory Management 28 UNIT – 4 Memory Allocation Techniques Memory allocation techniques are algorithms that satisfy the memory needs of a process: They decide which hole from the list of free holes must be allocated to the process. Thus it is also known as partition selection algorithms. There are primarily three techniques for memory allocation – First-fit Allocation – Best-fit Allocation – Worst-fit allocation UNIT - 4 Memory Management 29 UNIT – 4 Memory Allocation Techniques First-Fit Allocation – This algorithm searches the list of free holes and allocates the first hole in the list that is big enough to accommodate the desired process. – Searching is stopped when it finds the first fit hole. Next -fit Allocation – Searching is resumed from that location. The first hole is counted from this last location. In this case, it become the next-fit allocation. – First - Fit allocation does not take care of the memory wastage. UNIT - 4 Memory Management 30 UNIT – 4 Memory Allocation Techniques Best – Fit Allocation – This algorithm takes care of memory storage and searches the list, by comparing memory size of the process to be allocated with that of free holes in the list. – The smalled hole that is big enoughto accommodate the process is allocated. – It is better interm of memory of wastege but it incure cost of searching. UNIT - 4 Memory Management 31 UNIT – 4 Memory Allocation Techniques Worst– Fit Allocation – This algorithm is just reverse of the best-fit algorithm. – It search the list for the largest hole. – It is not good algorithm, in terms of memory, but it may be help ful in dynamic partitioning. UNIT - 4 Memory Management 32 UNIT – 4 Memory Allocation Techniques Example : Consider the memory allocation scenario as next slide. Allocate memory for additional requests of 4k and 10k (in this order). Compare the memory allocation, using – First – fit Allocation – Best – fit Allocation – Worst – fit Allocations UNIT - 4 Memory Management 33 UNIT – 4 Memory Allocation Techniques Problem : UNIT - 4 Memory Management 34 UNIT – 4 Paging Concept The first non-contiguous memory allocation method is paging. In this memory is divided into equal size partitions. The partitions are relatively smaller, compared to the contiguous method. They are known as frames. The logical memory of a process is also divided into small chunks or blocks of the same size as frame. These chunks are called pages of a process. Pagining is a logical concept that divides the logical address space of a process into fixed size pages, and is implemented in physical memory through frame. UNIT - 4 Memory Management 35 UNIT – 4 Paging Concept example There are 10 free frames in the memory. There are 4 processes P1, P2 ,P3, P4 consisting of 3, 4, 2, 5 pages respectively. For P1 (fig 10.15(b)) For P2 (fig 10.15(c)) For P3 (fig 10.15(d)) Now only one frame is free in the memory, where P4 required 5 frames. After some time P2 finishes its exectuion and therefore, release memory. These five frames, through non contiguous are alocated to P4 (fig 10.15(e ,f)). UNIT - 4 Memory Management 36 UNIT – 4 Paging Concept example Suppose after some time P1 release page 1, P4 release page 2 and P3 releases page 1 (fig 10.15(g)) Now P5 is introduced in the system with 5 pages, but only 3 pages to be accommodated in the memory (fig 10.15(h)) UNIT - 4 Memory Management 37 UNIT – 4 Paging Concept example UNIT - 4 Memory Management 38 UNIT – 4 Paging Concept example UNIT - 4 Memory Management 39 UNIT – 4 Paging Concept example A program's logical memory has been divided into 5 pages and these pages are allocated frames 2, 6,3, 7 and 5. Show the mapping of logical memory to physical memory. UNIT - 4 Memory Management 40 UNIT – 4 Paging Concept example UNIT - 4 Memory Management 41 UNIT – 4 Segmentation A programmer writes programs not in terms of pages, but modules, to reduce the problem complexity. There may be modules : main program, procedures, stacks, data etc. It would be better if memory management is also implemented in terms of these modules. Segmentation is a memory management technique that supports the concept of modules. The modules in this technique are called segments. The segements are logical divisions of a program, and they may be of different sizes, where as pages in the paging concept are physical divisions of program and are of equal size. UNIT - 4 Memory Management 42 UNIT – 4 Segmentation UNIT - 4 Memory Management 43 UNIT – 4 Segmentation Segementation has two advantages: – Segementation has logical memory which is closer to a programmers way of thinking – The segmentation need not be of the same size as compared to pages. UNIT - 4 Memory Management 44 UNIT – 4 Segmentation Segementation logical addreses: – Logical address of segement has two part The segment name (or Segment Number) Its offset – It has three major segment Code segment Data segment Stack segment UNIT - 4 Memory Management 45 UNIT – 4 Segmentation To convernt logical address of segment in to physical address use Segment tabel. Example : Segment Number Length/ limit Base Address 0 200 4100 1 700 1000 2 400 3700 3 900 1800 4 1000 2700 UNIT - 4 Memory Management 46 UNIT – 4 Segmentation UNIT - 4 Memory Management 47 UNIT – 4 Virtual Memory Virtual memory is used when process size is too large to fit in the real memory, therefore virtual memory is created. In virtual memory, combined approach of paging and segmentation is used. Virtual memory implementaion is complex as compared with real memory. It needs the assistance of hardware support known as paging hardware. Also OS have a module known as virtual memory handler (VM Handler). UNIT - 4 Memory Management 48 UNIT – 4 Need of Virtual Memory Paging and segmentation are two basic memory management techniques that require an entire process to reside in the main memory before its exectution. The increase in the degree of multi-programming means that more number of processes should be accommodated in the memory. But the degree of multi programming is limited with the size of the memory. This limitaiton may lead to several problems. UNIT - 4 Memory Management 49 UNIT – 4 Overlay The first solution was in the form of Overlay, years ago. An overlay is a portion of a process. A program is first divided into many overlays and store in the disk. A program containing overlaysis called an overlay structure program. This program consists of a set of overlay and a permanently resident portion known a root. As the root executes, the overlays are loaded as and whenever required. UNIT - 4 Memory Management 50 UNIT – 4 Overlay The required overlays are swapped in the memory and later on swapped out when the memory is full. Moreover, today, overlay is an obsolete technique. UNIT - 4 Memory Management 51 UNIT – 4 Virtual Memory Due to Overlay is obsolete, it gives rise to the concept of virtual memory in modern system. Virtual memory is a method that manage the exceded size of larger processes as compared to the available space in the memory. It means the degree of multi-programming can be increased without worrying about the size of the memory. UNIT - 4 Memory Management 52 UNIT – 4 Virtual Memory – Demand Paging VM system can be implemented using either Paging or segmentation. In Demand Paging, only pages that are needed at an instant of the time of execution are loaded. The benefits is that some pages, corresponding to some exception – handling or error- handling code, which may not be executed, are not loaded. It results in efficient utilization of memory and efficient execution. UNIT - 4 Memory Management 53 UNIT – 4 Virtual Memory – Demand Paging Demand Paging is same as Swapping. Except that an entire process is not swapped in or swapped out. Here, a Lazy swapper is used that loads only those pages that are needed. So here insted of “swap -in” is called “page-in” And “swap – out” is called “page-out” UNIT - 4 Memory Management 54 UNIT – 4 Virtual Memory – Demand Paging UNIT - 4 Memory Management 55 UNIT – 4 Virtual Memory – Demand Paging Demand Paging has some issues – (1) how to recognize whether a page is present in the memory. The page table with valid-invalid bit can be used for this purpose. Valid bit -> “1” page is memory at the time Invalid bit -> “0” page is either not valid or not present in the memory. UNIT - 4 Memory Management 56 UNIT – 4 Virtual Memory – Demand Paging Demand Paging has some issues – (2) it is a situation when a process execution does not get a page in the memory. A situation will occur in demand paging when the page referenced is not present in the memory. This is known as a Page fault. UNIT - 4 Memory Management 57 UNIT – 4 Virtual Memory – Page Replacement Algorithms When a page fault occurs during the execution of a process, a page needs to be paged into the memory from the disk. However, it may be the case that there is no free frame in the memory. In such case, an already existing page should be replaced so that there is room for a page that needs to be paged. This is known as a page replacement. If the page replaced by a random approach, then it may affect the performance. Thus, instead of replacing any page, the use of pages in the memory is to be observed and a page should be replaced such that effect on performance is the least. UNIT - 4 Memory Management 58 UNIT – 4 Virtual Memory – Page Replacement Algorithms The page replacement increases the overhead because there are two page transfer – Page – in & Page – out This over heads can reduced if it is known whether a page has been modified. Any instance of time it is not necessary every pages is modified and also some page are read only. In this case simply be over written by another page becase its copy is already on the disk. So one page – transfer time can be reduced. This is implemented by including M – bit or Dirty bit with each page. UNIT - 4 Memory Management 59 UNIT – 4 Virtual Memory – Page Replacement Algorithms UNIT - 4 Memory Management 60 UNIT – 4 Virtual Memory – Page Replacement Algorithms If the page is modified, then need to implement the page replacement algorithms. A page replacement algorithms must satisfy the following requirements: – The algorithms must not replace a page that many be referenced in the near future. It is known as non- interference with the program's locality of reference. – The PFR should not increase with an increase in the size of the memory. UNIT - 4 Memory Management 61 UNIT – 4 Virtual Memory – Page Replacement Algorithms Types of Page Replacement Algorithms – FIFO (First in First out Page Replacement Algo) – LRU (Least Recently Used Page – Replacement Algo) UNIT - 4 Memory Management 62 UNIT – 4 FIFO Page Replacement Algorithm According to FIFO, the oldest page among all the pages in the memory is chosen as the victim. All the page in the memory in a FIFO queue. The page at the head of the queue will be page – out first and a new page will be inserted at the tail of the queue. UNIT - 4 Memory Management 63 UNIT – 4 FIFO Page Replacement Algorithm UNIT - 4 Memory Management 64 UNIT – 4 LRU Page Replacement Algorithm In LRU, a page the that has not been referenced for a long time in the past may not referenced for a long time in the future either. LRU page – replacement algorithm replaces a page that has not been used for the longest period of time in the past. UNIT - 4 Memory Management 65 UNIT – 4 LRU Page Replacement Algorithm UNIT - 4 Memory Management 66 UNIT – 4 LRU Page Replacement Algorithm Stack Implementation – To implement LRU, a linked list of all the pages in the memory can be maintained. – The list can be structured as a stack such that whenever a page is referenced, it is placed at the top of the stack. – This way, the most recently used page will always be at the top and consequently, the least recently used page will be at the bottom of the stack. – This implementation requires removing one entry from the middle and placing it at the top of the stack. – The stack needs to be updated with every memory reference, which incurs a cost. UNIT - 4 Memory Management 67 UNIT – 4 LRU Page Replacement Algorithm UNIT - 4 Memory Management 68 UNIT 4 Completed UNIT - 4 Memory Management 70