Previously I had written a review on "Computer Networks" by Tanenbaum and Wetherall. I had no exposure to Kurose & Ross, so I thought it was the best around. I was incorrect. Don't get me wrong, the Tanenbaum book is still quite good. But I am now using Kurose & Ross for my networking class, and it is far better for an intuitive understanding of networking.
What makes Kurose & Ross better for a first course in networking? It reads quite well (except for Chapter 5, I think the editor forgot to take a look at that one), which is an incredible achievement for a networking book. The flow of the book is much better, as is the approach to explanations. They approach their explanations of networking as though you are a person who has no exposure to networking. I do not think the Tanenbaum book is quite so careful. What does this mean in terms of the book? K&R uses many analogies to help you understand. It also has an entire section devoted to everything that happens, step by step, when a user requests a webpage. DHCP, ARP, TCP, HTTP, etc. This is great for the big picture.
Also, K&R motivates the materials by explaining the dilemma they faced at the time, what the conceptual design considerations are, and then you are primed for the explanation of the actual protocols. This is vastly helpful. Another point for K&R is the top down approach. It really provide a significant improvement for the average student.
Lastly, I think the scope and content of K&R is better for a first course in networking. They are quite detailed, yet they leave out other things that are extraneous to our current understanding of networking (e.g. Shannon's limit and Nyquist's Theorem -- the way my networking prof put it, those have been settled debates for many years, and don't really affect the study of networking today). Which lends itself to my final point: the math used in K&R is WAY more practical and intuitive. Propagation, Transmission, and Queueing delay, queueing efficiency, the difference in time between persistent and non-persistent HTTP, etc. Tanenbaum doesn't really have the unified picture in this regard, and some of the math in it is very unrealistic for a first course.
All in all, an incredible networking text.