Customer Reviews

Reliable Distributed Systems: Technologies, Web Services, and Applications

5 star:		(1)
4 star:		(1)
3 star:		(1)
2 star:		(1)
1 star:		(0)

 

 

This book is neither a book in distributed applications neither a book about web services. It is true thta the author makes an introduction of both and in the case of the web services he takes a look to the latest standards related to reliability.
From my point of view, the book talks about how to enhace the reliability of a distributed software application using static and dynamic groups.
The author thinks that solving the life cycle problem of the nodes in a distributed aplication (create and destroy nodes, joint and leave groups, heart beat, and so on) is the necessary step towards a reliable distributed application.
I think it is an interesting point, however I am not convinced at all. Perhaps, I should read it again.

I really recommend it. For a more straight forward introduction to the subject perhaps is better to start with the classics Tanenbaum or Colouris books.

If you tried Nancy Lynch's "Distributed Algorithms" and only care about web apps, this might be the book for you. It's readable, useful for people developing web back ends, and up to date. However, it doesn't try to cover the complexities of general fault-tolerant distributed systems. Besides Lynch's classic text, I recommend Gray and Reuter's "Transaction Processing," a seminal text that should give you a clear idea of how to compose arbitrarily reliable systems out of unreliable components.

I have been doing a bit research of my own in this field, and you really have to be there to appreciate the value of this book, which by far provides the most comprehensive review of this "old" but yet challenging field, i.e. how to build reliable distributed systems that actually perform, among all the related topics.

The book covered the subject in both depth and breadth, and in a wide time and scope range. It clarified a lot this confusing area, for both researchers and professionals.

Birman gives a lucid explanation of various ideas involved in making a distributed computing system. Some have been around for years, like CORBA. Yet he advises, correctly, that its industry implementations have many difficulties regarding ease of interoperability. Which has greatly constrained CORBA's success.

A lot of space in the book goes over Web Services. These take a radically different approach to making a distributed system. Where there is an emphasis on structured messages that are not in binary, like CORBA. And where the messages are written in XML. Still too early, but Web Services might well overtake CORBA.

Peer to peer networks also get attention in the book. Gnutella, Napster, Chord and others. Yet some like Chord, have mostly been theoretical formulations. Interesting potential, but little market acceptance. Surprisingly, for a book that came out in 2005, there is no mention of BitTorrent. Yet in 2006, BitTorrent is perhaps the most popular p2p network.