Various kinds of distributed systems have been proposed and built in the past few years. These range from loosely coupled systems connected by low-bandwith, wide-area public data networks, through microcomputer + memory nodes connected by fast local networks; to tightly coupled shared memory systems. This book is about the design and analysis of the second type of system: independent machines within a room or building connected by a local network.
The book is an outgrowth of a five-day tutorial course which has been stretched to fill a book. It covers a broad area, but is, as a consequence, rather superficial. To a student who has already had a good course on operating systems, much of the material will seem straightforward, if not simpleminded. On the positive side, the material is easy to read, even for readers with little background in this area.
The 12 chapters are divided into four parts: Introduction, MPS Building Blocks, Design Tools and Techniques, and Example Systems. The Introduction categorizes distributed systems and describes their advantages and disadvantages. The Building Block part deals with hardware and software, with most of the emphasis on the former (only one 14-page chapter of fairly sketchy material on distributed operating systems is provided, for example). Part 3 is about calculating response time and mean time to failure. Part 4 summarizes some example systems, such as the Coast Guard’s Vessel Tracking System.
In short, the book is best suited for a quick introduction to the subject for the reader who might otherwise take a five-day course to get acquainted with distributed systems.