In this book, we are studying the correct development of distributed programs by means of various examples. So far we have done this in Chapter 4 (file transfer protocol) and in Chapter 6 (bounded retransmission protocol). In later chapters, we shall also study some distributed program developments: leader election on a ring-shaped network in Chapter 10, synchronizing processes on a tree in Chapter 11, routing algorithm in Chapter 12, leader election on a connected network in Chapter 13. We shall also study the correct development of sequential programs in Chapter 15. In this chapter, we shall study another kind of execution paradigm, namely that of concurrent programs.

Comparing distributed and concurrent programs

The distinction between sequential and distributed programs must be clear. But the one between distributed and concurrent ones might be less obvious. Here are the main differences which we consider between the two.

Distributed programs

In the case of distributed programs, the entire algorithm is performed by various agents executing some sequential programs (sometimes the same one) on different computers. But, at the same time, these agents are supposed to cooperate in order to achieve together a well-defined goal, which is the purpose of the algorithm.

This cooperation could be made easy by having a centralized agency, the role of which would be to schedule the various participating agents. But we suppose that such an agency does not exist.