algorithms

Professor's Nancy Lynch's "Distributed Algorithms" is a definite reference for theoretical treatments of many hard problems in distributed computing. It is a textbook, but written in such a clear style that makes it almost a pleasure read. Rarely have I seen something like that! The book has a right proportion of theoretical proofs, practical applications, philosophical appreciation of the problems, research questions, examples and study points.

"Distributed Algorithms" has 3 main parts - synchronous, asynchronous and partially synchronous network algorisms. Each part describes consensus resolution, mutual exclusion, resource allocation, leader election, termination detection and failure detection as main problems in distributed computing theory. Lynch has done a masterful job of leading us from simple to complex, from theoretically solvable to practically intractable problems.

For a practitioner of computer science, who is not necessarily involved in fundamental research, this book gives a clear appreciation of problems of 2PC, resource management, failure profiles in faulty and noisy networks, optimization and fault management in distributed networks. All those things are foundations of databases, network computing and enterprise scalability. It also helped me greatly in estimating the best and worst case boundaries in certain practical distributed system optimization problems.

At MIT, I took the course 6.852 Distributed Algorithms under Professor Nancy Lynch. It was an excellence course. This books is based on her lecture notes. Before this book, there is really no book that covers these material in rigorious and consistent matter. One usually have to read the actual published academic papers. Because different author may use different notations or models, sometimes it is hard to see the whole picture. This book shows exactly that. The algorithms are presented in a consistent notation, and the models and the assumptions all the explicit, clear and consistent. However, Professor Lynch's lecture style can get really dry and boring, sometimes I can't help felling asleep because the class was so early in the morning. For that, I decided to give it only 4 stars.

This book is in the same class as "Discrete mathematics" by Knuth and others. Important topic, extensive coverage, good English, zero vendor's propaganda. Super. An unexpected gift from up above (after struggling with reams of MS's (dis) information <g>.) I am working on something distributed and ran into this book accidentally, while browsing in a bookstore--I'm glad I did. Btw, it's a few bucks cheaper in B&N store (here goes my review <g>.)