SYNCHRONIZATION TECHNIQUES FOR DISTRIBUTED SYSTEMS - AN OVERVIEW

This paper discusses various synchronization mechanisms and primitives which have been introduced for distributed systems and their concurrent programming languages. Distributed systems originated in the need to increase computing power through the technique of breaking a large problem into a number of smaller components (processes), running these smaller processes on separate processors, and recombining their results to solve the original larger problem. The main issues of such techniques are the splitting, interaction and recombining of these smaller processes. Each of these issues has two major components: mutual exclusion and synchronization between resources and/or processes. Generally, synchronization includes the issue of mutual exclusion. This paper provides an overview for individuals who are not overly familiar with distributed systems. The paper will then focus on synchronization primitives, and a few proposed implementations of languages utilizing these primitives.