Computer Networking: A Top-Down Approach

This review compares the following four books: Computer Networks by Peterson and Davie (P & D) Computer Networks by Tanenbaum Computer Networks by Comer / Internetworking with TCP/IP Computer Networking by Kurose and Ross (K & R) By far the best book in the list is "Computer Networking" by Kurose and Ross. This book covers all of the essential material that is in the other books but manages to do so in a relevant and entertaining way. This book is very up to date as seen by the release of the 5th Ed when the 4th Ed is barely two years old. There are lots of practical exercises using wireshark and the companion website is actually useful and relevant. The attitude of this book with regard to teaching networking concepts could be summed up as "try it out and see for yourself". One interesting thing to note is that the socket programming example are all in Java. Next up is the Peterson and Davie book which covers everything that Kurose and Ross discuss but is slightly more mathematical in how it goes about things. There are a lot more numerical examples and defining of formulas in this book which is fine by me and in no way detracts from the book. Also the socket programming examples are in C which is a little more traditional. The points where this text loses ground to K & R is that it doesn't have the practical application exercises that K & R has and it also doesn't extend the basic networking theory that is covered to modern protocols like K & R. The two Comer books come next. Comer's "Computer Networks" book is probably the most introductory book out of this whole list and is more of a survey of networking topics that doesn't cover anything in any real depth. Still, this is an excellent book in that it is a quick clear read that is very lucid in its explanations and you can't help feeling that you understand everything that is covered in the book. Comer's TCP/IP book is the equivalent of the other authors' computer network books and in that respect it is pretty average. It covers all of the relevant material and in a manner which is more than readable but that is all. There is nothing exceptional about the book which stands out from the rest. Last comes Tanenbaum's book from the author who is probably most famous for his OS books. This is probably the most technical and detailed of the books with lots of sample C code belying is experience with operating systems and their network stack code. The weak point of this book is that all of the code and technical minutia might prevent the reader from seeing the forest for the trees. Unless you are trying to learn how to program your own network stack for a Unix/Linux system, then I would get either the K & R book or the P & D book to learn networking for the first time. This book would best be served as a reference in which case the technical nature of the book becomes a benefit rather than detracting from the text.

"Computer Networking" by Kurose and Ross is an excellent introduction to the topic of computer networking as it relates to the largest and arguably most important network of them all: the internet. As stated in the title, the authors take a top-down approach that, in my opinion, is very well suited to CS or SE majors who are typically already familiar with aspects of the application and transport layers, but who often have had little or no training (or interest) at the physical layer. In this way Kurose's text distinguishes itself from other popular texts such as those by Stallings ("Data and Computer Communications") or Tanenbaum ("Computer Networks"). There are a total of nine chapters, the first of which paints the big picture and lays out the plan for the book. The book presents a analysis of each layer of the 5-layer TCP/IP model. The decision to use the TCP/IP model over the OSI model is appropriate considering the authors' intentions of focusing on the internet. To accomplish their goal the authors next provide chapters on the application, transport, network and data-link layers respectively. The coverage of the application and transport layers are where the book seperates itself from both Tanenbaum and (especially) Stallings. In addition to providing the theory Kurose and Ross also inject practical demonstrations such as introducing the reader to socket programming in Java. The coverage of the network and datalink layers is somewhat more standard, but still very good and complete. Following these are chapters on wireless and mobile networks, multi-media networking, network security, and a small chapter on network management. Coverage of the physical layer is spread throughout all of the chapters in the book and presented on a "need-to-know" basis. Each chapter presents ample details about the protocols, technologies, algorithms and standards that are used by each layer as it relates to the internet. All of these chapters are quite long (between 70-120 pages) which may give readers the impression that the text drags on. It would be wise for readers to treat the major sub-headings in each chapter as single entities, rather than attempting to plow through each chapter in a single sitting. The style of the writing would be best described as clear and accessible. The authors make use of analogy and repetition to drive home the content, particularly in the first three chapters. Students who are new to the subject and have not had much experience thinking about things such as protocols and algorithms will likely appreciate this approach. Like any book, this book will be useful only to a subset of people who want to study computer networks. I think this book provides the best introduction for CS/SE students who have not had much exposure to network theory and are looking for a practical introduction to the subject. For CS/SE students who feel they want a more formal, terse, and dense introduction to the subject, they sh

It's hard to believe that a "technical" book can be so captivating - I could hardly put it down once I start reading it. My only complaint is the price. However, look out for those who sell "Softcover Intl edition with exact same content". I made the mistake of trying to save some money and bought one, and the quality of the book was simply terrible - flimsy paper, ink shows through, and all the figures were in poor black-and-white copy (instead of the nice shades of gray with blue highlight). On the back of my book it says "For sale only in Indea...".

I'm a graduate student in Comp Sci, and I recently had the opportunity to take a networking class again just to refresh my basic knowledge (my dissertation topic isn't related to networks). It was a pleasant surprise for me that the class utilised this textbook. I have been at the university level in CS for 8+ years (grad + undergrad), and this book is *by far* the best introductory computer science textbook I have ever read in any CS subject.The book is very well-written and extremely interesting to read. I was never bored in any chapter. Kurose and Ross are knowledgable experts in their field, and their exposition of the material is fantastic. Unlike Tanenbaum's book, they start at the application layer and move down. IMHO, this is a far better pedagogical strategy, because young students these days already have an excellent layer-5 understanding thanks to daily interaction with HTTP, IM, P2P file sharing, etc. If I remember correctly from my undergrad days, my own experience in a bottom-up approach, starting at the physical layer, really put me to sleep and put me off from networking. That's a shame, because networking is a really exciting field.The best parts of the book are the breadth, thorough use of real-world topics, and the illustrations. In fact, the diagrams and illustrations are just plain great. Most technical writers often rely too much on the written word. Here, the authors augment almost every pair of pages with an illustration; this is simply remarkable. The explanations of fundamental topics (such as packet-switching, DNS, TCP congestion control, IP routing, and ethernet) are *extremely* clear. More advanced topics are very up-to-date, covering cutting-edge subjects such as P2P, CDNs, security, NATs, 802.11, RTP, etc. I have not found a better introductory explanation of CDNs anywhere else. Although networking does have a lot of math in various areas, this introductory book does not get too much in detail in mathematical discourse, making this book very readable. That's a fine approach in my opinion, as a deep mathematical analysis of various topics is best left for grad school or a professional job.The authors' academic background really shows. Every topic is filled with citations/references to other work. This is great, because this book is just an introductory book with wide breadth but is otherwise lacking in significant depth. The interested reader (future grad student or network engineer?) can easily follow up on any topic he/she likes thanks to the exhaustive list of references.As if all of that were not enough, there is an accompanying website that has interactive Java applets demonstrating various topics as well as a set of Powerpoint slides for download. Furthermore, I enjoyed the interchapter dialogues with various famous researchers in the field.All in all, this is an outstanding book for the undergrad level, and I expect this would be a great book for professionals who want to have a firm grasp on networking funda

I've taught Park University's Computer Networking course four times now and I've used four different textbooks. I'm very happy with my most recent choice - Kurose/Ross's Computer Networking. The Kurose/Ross textbook is everything I've been looking for.Our computer networking course is an overview of networking. Many of the textbooks on the market are overly mathematical or overly business-oriented. This one is just right. Personally, I love Tanenbaum's networking textbook, but I have to admit that this one is better suited for an introductory networking course.I thought that the PowerPoint slides were pretty good, but I still didn't use them much because I'm ultra-picky about my lecture presentation and I thought that I needed to write my own slides for added clarity. Positive attributes:1. Emphasis on the Internet and the Internet protocol stack. As the authors point out, that makes for a more relevant model than the old OSI model. Students can more easily get a handle on the Internet protocol stack model.2. Great website.3. Well written.4. Great content. Lots of detail when you need it, but it introduces difficult concepts in a manner that is not overwhelming.5. Good homework problems. I haven't tried the programming excercises because my students didn't know Java, but next time I teach the course, the students will know Java and I plan to use the programming assignments.6. Great response from the authors. I enquired about a typo and Keith Ross replied the same day with appropriate help.