Differences between processes and threads in operating systems.

A process is a program in execution, whereas a thread is a path of execution within a process. Processes are generally used to execute large, ‘heavyweight’ jobs such as running different applications, while threads are used to carry out much smaller or ‘lightweight’ jobs such as auto saving a document in a program, downloading files, etc. Whenever we double-click an executable file such as Paint, for instance, the CPU starts a process and the process starts its primary thread.

Fig 1.1- A thread within a process

Each process runs in a separate address space in the CPU. But threads, on the other hand, may share address space with other threads within the same process. This sharing of space facilitates communication between them. Therefore, Switching between threads is much simpler and faster than switching between processes

Threads also have a great degree of control over other threads of the same process. Processes only have control over child processes.

Also, any changes made to the main thread may affect the behavior of other threads within the same process. However, changes made to the parent processes do not affect the child processes.

Considering the similarities, a thread can do anything that a process can do. They are both able to reproduce (Processes create child processes and threads create more threads), and they both have ID, base priority, execution time and exit status as attributes.

                   

Fig 1.2- Life cycle of a process

To understand the actual differences between processes and threads, let us consider a real-life example. Suppose you are driving from city A to city B and you have had only three hours of sleep the night before after a terribly hectic day. Imagine yourself as the CPU and driving to city B as a process. While driving you may have to do a number of smaller tasks such as checking the map on the GPS, making sure that you don’t exceed speed limits, making sure you don’t fall asleep (trust me, it happens!), etc. These little tasks are actually threads of your process because collectively they serve to help in its execution. If any thread fails to execute properly, the entire process may be disrupted. (For instance, imagine what would happen if you actually fell asleep!)

Now let’s assume that there had been a storm the night before. After travelling some distance you see that a toppled over tree is blocking the road ahead. You cannot move any further unless the tree is removed. Therefore, your current process has to be stopped. Now you’ve got a brand new process in your hands- getting the tree out of the way.

So you gather people from a nearby village, which can be thought of as a thread to your new process, and shove the tree out of the way. Then you go back to your car and continue with your journey.

Fig 1.3- Check out my drawing skills!!

Let us take another example. Suppose you are cooking pasta in your kitchen. Once again, imagine that you are the CPU and cooking pasta is your process. Your threads this time will be rolling and cutting the pasta dough, preparing the sauce, grating cheese, etc. All of a sudden from the corner of your eye you notice something moving. You turn towards that direction and to your utter horror you see a huge rat! Now you cannot cook with the knowledge that there is a huge, nasty rat at large in your kitchen. So you stop doing whatever you were doing and set about accomplishing a new task, or a new process- Getting rid of the nasty little rodent. So you cut a little slice of cheese, take a stick and wait for the rat in ambush. All of these are your threads. Finally when the rat does show up, you try to hit it and miss terribly, but you do succeed in scaring it away from your kitchen. Satisfied, you peacefully get back to your cooking.

Considering threads in windows and linux

In Linux, there is no distinction between processes and threads; instead it simply uses the term task. On the contrary, in windows there are considerable differences between these two as has been mentioned above.

The limited distinction between threads and processes in Linux exists because the entire context of the process is not present in the main process data structure. Instead the context is contained in independent data structures. The process data structure simply contains pointers to these separate structures. In windows, on the other hand, all the data vital to the process are contained exclusively within the process itself.

In Linux, the scheduler represents a higher priority of execution by a lower number, whereas in Windows it is represented by a higher number.








References
-Modern Operating Systems by Andrew S. Tanenbaum and Albert S. Woodhull
-Operating System Principles by Abraham Silberschatz, Peter Baer Galvin and Greg Gagne
-http://www.programmerinterview.com/index.php/operating-systems/thread-vs-process
-www.wikipedia.com
-http://www.programmerinterview.com/index.php/operating-systems/thread-vs-process
-http://rossano.pro.br/fatec/cursos/soi/Linux.pdf
-http://avellano.fis.usal.es/~lalonso/amp_inf/windows.pdf
-http://codesam.blogspot.com/2011/03/introduction-to-thread-in-java-part-2.html