9.1.1. Introduction to Virtual Memory

In this module we will start our discussion with Virtual Memory. In the previous modules we had studied caches which provide fast access to a processes code and data. In a similar way we will now study a technique which allows the main memory to be used as a cache for the secondary storage. And this mechanism is jointly managed by both the OS, and the CPU hardware.

This first chunk introduces the concept of virtual memory in computer organization. Virtual memory is a technique that acts like a cache for secondary storage (like hard drives), allowing faster access to data and code. It is designed to make the main memory (RAM) work more efficiently by managing how data is stored and accessed. This system is managed by both the operating system (OS) and the hardware of the CPU, ensuring successful communication and data retrieval.

Think of virtual memory like having a filing cabinet (the main memory) where only the most frequently used files (data and code from active processes) are stored, while the less used files (from secondary storage) are kept in a larger storage room. The filing cabinet allows quick access to important documents, just as virtual memory allows quicker access to processes running on the computer.

So, virtual memory allows active portions of multiple processes to be concurrently resident in the main memory. So, as we know that multiple processes must exist together in the main memory for being executed.

Virtual memory enables the simultaneous presence of portions of multiple processes in main memory. This capability is crucial because it allows the computer to run several applications at once without needing each one to fully occupy the main memory at all times. When a program is executed, its active parts are stored in main memory, while inactive parts can remain in the secondary storage, making room for other processes.

Imagine a restaurant kitchen where various chefs are preparing different dishes at the same time. Each chef can only work on certain elements of their dish at any given time, while the rest can be kept in a pantry (the secondary storage). Virtual memory in a computer works similarly by allowing multiple processes (chefs) to share the main memory (kitchen) effectively without overcrowding.

The virtual addresses that are generated by the CPU must be mapped to physical addresses. So, I generate a virtual address there will be a mapping process, the memory management unit will map this virtual address into a physical address.

When a program is executed, the CPU generates virtual addresses for data access. However, those virtual addresses do not directly correspond to physical memory locations. Instead, the Memory Management Unit (MMU) plays a vital role in translating these virtual addresses into physical addresses. This translation is crucial for ensuring that data is retrieved from the correct location in the main memory.

Consider a mail delivery service that uses postal codes. The virtual address is like an address with just a street name (virtual) that requires mapping to the actual building location (physical). The postal service (MMU) takes the street name and finds the exact location to deliver mail (access data) in the city.

Now this translation process also enforces protection of a programs physical address space from other programs. This protection is enforced by the translation process.

The translation from virtual addresses to physical addresses provides not only mapping but also protection. Each process is prevented from accessing the memory of other processes. This safety ensures that programs do not interfere with each other, maintaining the integrity and isolation of the operating system and its processes.

Imagine a library with separate study rooms for different groups of students. Each group can focus on its work without other groups intruding, similar to how virtual memory prevents one program from affecting another. Just as staff ensures students only access their designated rooms (memory), virtual memory systems protect processes using separate address spaces.

Now, the translation process also allows a single user program to exceed the size of the main memory because virtual address space and physical address spaces do not have a one to one mapping.

One significant advantage of virtual memory is that it allows programs to use more memory than what is physically available in main memory. Since the sizes of the virtual address space and physical address space are independent, a program can utilize a larger virtual memory while only needing the necessary portions in the physical memory at any time.

Think of a digital library (virtual memory) that offers access to a vast collection of eBooks much larger than the library's physical shelf space (physical memory). Users can browse and read from a massive selection without needing to store all the books in a physical space, retrieving just what they need when they need it.

Now, this virtual memory translation miss is called a page fault; that means, during translation I have my CPU asked for data corresponding to a virtual address...

A page fault occurs when a program attempts to access data that is not currently in physical memory. When this happens, the operating system must retrieve the required data from the secondary storage to load it into main memory. This process helps maintain the illusion of a large, available memory space.

Imagine trying to access an important document stored in a filing cabinet, but realizing it’s currently in a storage room (secondary storage). You’ll have to go get it before you can continue working. This delay in fetching the document is similar to a page fault in virtual memory.

--