This book is definitely dated. That's no surprise in a technology book. However, the basics behind network programming are still sound and worthwhile; and while the vocabulary has changed the principles in the book underlie Voldemort, Cassandra, MapReduce, Thrift, Protocol Buffers and more.
Unfortunately the book predates the Eight Fallacies of Distributed Computing ( http://en.wikipedia.org/wiki/Fallacie... ). Read those first, and refer to them several times while reading the book so that you can take some of the objectives with a large enough grain of salt.
Old, outdated and obsolete, which is of course inevitable with pretty much every technology book. Nonetheless, it was pretty good in its time and still may be a good historical or foundational resource for some people.
I read this book way back in 1997 and thought it was an excellent book which summed up the state of the art at the time and gave valuable pointers to ongoing research and more detailed work.
It included topics on synchronizing logical clocks, distributed election algorithms (which I thought lacked work by García-Molina), distributed shared memory, threads, scheduling processors, fault tolerance, real-time distributed systems, distributed file systems and case studies on Amoeba, Mach, Chorus and DCE.
Andrew Tanenbaum is one of the great textbook writers and researchers on distributed operating systems. As I write (2012) his most recent textbooks are the fifth edition of Computer Networks (2010) and the third edition of Modern Operating Systems(2007). His book on Computer Organization is about to come out.