Computer Networks

First of all, the fourth edition was published in 2002, so all reviews prior to that date are about a previous edition of this book. This is a classic textbook on computer networking from an academic viewpoint. Do not expect to ever be able to fix a specific network problem or become a CCNE by reading this book. However, doing either of those tasks rests on a firm foundation of the theory found in this book. From the beginning, the author points out that there is some confusion about what a computer network is - a collection of autonomous computers connected by a single technology. He then points out that actually neither the world wide web nor the internet are computer networks. The book goes on to explain networks in terms of a 5 layer system rather than the classic 7 layer OSI model, which is the same as in the previous edition. However, much has been added and much deleted based on the rapidly changing technology involved. For example, the chapter on the physical layer has been completely rewritten. The previous edition focused that chapter on ISDN, ATM and cellular radio. The current edition omits references to that technology and discusses the mobile telephone system and cable television instead. As would be expected, the other section of the book that had the biggest revision was the chapter on the application layer. Gone is the obsolete subject of USENET news, multimedia has changed completely, and the network security section now has its own chapter due to the importance that field has taken on. Finally, the chapter on further reading, which had good comments to go with the suggested reading, was always one of my favorites because it told you why you should read something in addition to showing you what to read, plus the bibliography is divided by network layer. Now, of course, the 3rd edition bibliography section reads like the roll call for a computer book museum. Thus, the new 4th edition chapter on suggested reading is a welcome update. Remember that this book is ultimately about good network design decisions as well as a tutorial on architectures, thus math is inevitable. In fact, most of the chapter-end problems involve mathematics. The math required is not complex, but it will require careful reasoning on your part, much like the numerical problems in Hennessy's "Computer Architecture: A Quantitative Approach". A good companion to this book is "Computer Networking : A Top-Down Approach Featuring the Internet (3rd Edition)" by Kurose. That book also uses a 5 layer model, but starts from the application layer and works its way down. Also, Kurose's book is more applied, with Java examples and programming labs and projects included. I highly recommend this book to anyone who wants a firm foundation in the theory of computer networking, as well as a survey of the current networking technologies and how they work.

Computer Networks are a wide and fastly growing subject. Finding a textbook that covers all of the topics in a detailed way is simply impossible. Perhaps for this reason good textbook authors have, in a probably implicit way, established two possible approaches: the Engineers' and the (mostly Software) Developers'. Once again Tanenbaum has done a great job with this book (and its updated-more-than-revised 4th edition), which takes the former approach.The book presents general issues and impacts (on technology as well on the society) of Computer Networks in the first chapter, and then move in a detailed exposition of the lower layers of a general network architecture (similar to the OSI one). The great value of the books stems from the clarity and thoroughness of the exposition. Indeed, it presents all of the most known technologies and algorithms (both today's and historical) from physical mediums to algorithms for routing, congestion and flow control and so on. Plenty of details are provided at the level of mathematical performance analysis for some algorithms like those presented in the Medium Access Sublayer chapter (e.g. ALOHA and CSMAs).The "tone" of prof. Tanenbaum is an added values as well. He rarely becomes boring and sometimes results hilarious in his comments of famous anecdotes that led to the born of this technology or that algorithm (have you ever heard how automatic phone calls switching was born ?). I never underestimate the value of an easy exposition, as sometimes studying is already hard enough to cope also with a overwhelmingly boring book.Enough for the lower layers/protocols so far. About the upper ones the book actually does not spend too much emphasis on network applications nor on the high level tools for building network applications (e.g. there are a very few pages for sockets, but no more). Indeed, this area is more properly in the competence of the second kind of books (Developer's) as noted at the beginning of the review. However, there's one (unsurprising but happy) exception: as already done in his "Modern Operating Systems, 2e", Tanenbaum has put a detailed and rigorous treatment of the Security issue (Network Security in this case).About editions, the third was already a very good book. Reasons for considering the fourth edition are the inclusion of updated technologies like ADSL, Bluetooth, Gigabit Ethernet, JavaScript, XHTML or XML, etc. More than this, however, technologies like fiber optic were on the wave of great improvements in 1996 when the third edition was published (and deformation due to day-night thermal excursions were not cited) so that now the treatment is more reliable (in terms of updates, not in technicalities).All in all, given that imho there's no serious "complete bible" (or the like) book on computer networks, this book is a full five-stars one if the Engineers' perspective is that of interest. If one is more interested in the Developers' perspective (take again the sockets example), then a

When I just started working as a programmer, I came across the 2nd edition international student reprint of this book, in red softcover, at a students' bookshop, and bought it. I treasure that copy in my library to this day - it gave me a solid background in computer networking that served me well in the past and is an asset that will serve me in the future.Previous reviews have mentioned Tanenbaum's excellent and concise writing style and the breadth of the material covered, which are two of the good qualities that made this book a classic.What wasnt mentioned and I consider important is :- The bottom up explanation of networking, starting with the physical layer, up through the OSI layers (link layer, network layer, transport layer) to the application layer. This structure makes it very clear how networks work, how they are designed, and why.- The well made balance between broad coverage of topics and the depth of coverage of each topic. After reading this book it's easy to start studying in depth any subtopic covered in this book (e.g. TCP/IP through Comers' books) as the reader knows how the subtopic fits in the big picture and knows enough to make a smooth start.Some previous reviewers asked about solutions to problem, so I'll note that this book does have booklet with problem solutions, which is apparently made available only to academical staff [ and people mistaken to be such :-) ].I highly recommend this book to anyone who has any type of network related job, students who study networking, and anyone who wants to study computer networking.

This is probably the most complete and in depth theoretical approach to Computer Networks book you can buy.The structure of this book is based mainly on the OSI Reference Model which is divided in 7 different layers. Each chapter in the book is dedicated to a layer, in this way, one can get an insight of a broad variety of subjects such as IEEE standards (Ethernet, Token Ring, etc,), Congestion control and routing Algorithms, network and transport layer protocols (IP, TCP and UDP) and many more topics.As a complement you should read Richard Stevens TCP/IP Illustrated, volume 1 for a more practical approach to TCP, UDP and IP.The best book for beginner, intermediate and advanced users.