Real Memory

In computers, memory is referred to as a mechanism with the ability to store data and information. Storing data within a computer can be done in a variety of ways across multiple device platforms. Desired data information can be stored permanently or even temporarily. Unquestioned in computing, memory management is the fundamental act of properly distributing the appropriate and most fitting portions of memory among programs. This is all possible due to a unit known as real memory.

Real memory deals with the actual hardware of a computer known as memory chips or commonly known as ram. Memory chips are physically apart of any computer after installation. Every single program is run through this piece of hardware which is physical memory. Still, as with any piece of hardware there comes constraints. Physical memory has a certain limit of resident set that can be loaded. Resident set refers to the portion of the process image that is actually in the real-memory. However, there is a way to bypass that constraint. By the use of virtual memory.

Now with virtual memory hardware constraints are broken and far exceeded. To put things into perspective virtual memory simulates to the computer to believe that it has more physical memory available, but in all actuality that extra portion of memory is all virtual. The exact process is done by separating the memory addresses which is used by a given process from the actual physical addresses. This split of processes grants great benefit to the efficiency of the available amount of memory in a system. This allows the system to multitask and as well as allow the system to run a far much larger program. Virtual memory is possible via paging or segmentation.

Despite the advantages presented...

... middle of paper ...

...there will be thrashing. In a multiprogramming level, process utilization rises but a situation is reached when the average resident set is inadequate and the page fault increases which result in the collapse of processor utilization. There are many ways to deal with the problem. The PPF algorithm implicitly uses load control. Only those processes that has large resident set is allowed to execute. Another policy is the 50% criterion, where the utilization of the paging device is kept at approximately 50%. Another approach is the clock page replacement algorithm. It uses a global scope that monitors the rate at which the pointer scans the circular buffer. If the rate is below a given lower threshold, few page faults are occurring and, or the number of frames scanned is small meaning that the resident pages are not being referenced and can be replaced easily.