Explain semaphore concept in multiprogramming

Answer 1

A semaphore is a service very often offered by real-time operating systems to allow programmers to perform one of two major functions: synchronize 2 tasks or control sharing of resources between 2 or more tasks. There are different kinds of semaphores, but one of the most common is called a binary semaphore. In this kind of semaphore, only 1 task at a time may "have" the semaphore. The simplest way to think of the binary semaphore might be to consider the semaphore like a baton and there is only one baton. Say if you want to allow only 1 person to speak at a time (consider people to be tasks), then you make a rule that only the person holding the baton may speak. A person wanting to speak will go and pick up the baton. If nobody else has it, he may speak. If someone else has it, he must wait until the baton is put down by whoever has it. This is the "sharing resources" way a semaphore is used. The task synchronization usage, similarly can be thought of using this baton. Just as in a relay race where one person cannot start until another passes the baton, a binary semaphore can be used to synchronize two tasks by giving one task the semaphore, having the second one wait for the semaphore, and then having the first give it to the second. The two tasks will be synchronized at the point the semaphore is passed from the first to the second just as a baton is passed from one runner to the next in a relay race. You can have semaphores other than binary semaphores, which are called counting semaphores. Counting semaphores are just binary semaphores generalized to more than the 2 values of 0 and 1, to some value n.