Part One. The Swap space
In order to work properly a computer depends on having an adequate amount of memory. Simply saying that there can never be enough. The more physical memory is installed the more costly it is. Mostly, the result is a clever compromise between costs and speed to access the memory cells.
To achieve this compromise UNIX/Linux systems combine two types of memory — physical memory (RAM), and swap space. Altogether this is called the virtual memory of a computing system. Physical memory is rather expensive but fast and accessible within nanoseconds. In contrast, swap memory is rather cheap, but slow, and accessible within milliseconds.
There exist a few reasons why swap memory is useful. First, sometimes single processes need more memory than the system physically owns and can provide more to the processes that demands it. As a result, all data that is kept in physical memory cannot be stored there any longer. Now, the swap space comes into play, and a selection of memory pages are transferred to the swap space to free physical memory.
Second, not all the data is needed in memory at the same time. That’s why less used memory pages are parked on swap space to have as much free physical memory available as possible. This method is named the Least Recently Used Page Replacement Algorithm (LRU) [1].