Browse Books

Reviewer: Antoine Joux

Computer security is a growing concern in our information-based society. The goal of this book is to present the state of the art in this field to a general audience, including computer security professionals and students. The book's organization is very similar to that of another book by Bishop [1]. A large part of the text, and of the examples and the exercises, is also common. As is its predecessor, the book is divided into several parts and many chapters. For some reason, the parts are not made explicit in the book, or in the table of contents. According to the table of contents of Art and science [1], the parts are: "Introduction," "Foundations," "Policy," "Implementation: Cryptography," "Implementation: Systems," "Assurance," "Special Topics," and "Practicum." A few extra chapters, with some background information, have been added at the end. This very similarity is the weak point of the book. The structure of a largely mathematical textbook, such as Art and science [1], is not suited for a broad audience. Just to give an example, this book's intended readers are probably much more familiar with mass operating systems, such as Windows. Thus, the large amount of Unix-based examples and exercises is probably missing the point. Similarly, the programmer-oriented issues, which offer nice examples of good and bad programming practices, are too technical to be of general interest. The style of writing and the focus of the book vary greatly from one part to the next. Some parts are captivating, and others are very formal, even though the mathematics have been omitted. The most interesting part is the discussion of various threats and attacks, especially chapter 19, on malicious logic. There, the sections on Trojan horses and viruses could have been expanded, to cover more recent examples. In particular, several recent Trojan horses and viruses make use of the preview capabilities of some mail readers. Knowledge of these threats could be very useful to readers. The practicum part is also interesting, however, the chosen example is probably too complicated, and it would have been nice to see a simpler example, focusing on the security of home computers. On the other hand, the parts about policy and assurance are formal, and hard to read. Worst of all is the exposition of cryptography, which is at best outdated, and often misrepresents the current knowledge in this field. Classical cryptosystems are not used, and should not be used, in computer security, so why present them in such detail__?__ A cryptosystem is not only a tuple of sets and functions, it should also satisfy security properties, such as indistinguishability from an idealized primitive. These properties have been developed into a nice theory of concrete provable security by Bellare and others. Similarly, plain textbook RSA is known to be weak, due to a large variety of multiplicative attacks. It would be nice to present the correct usage of the primitive, which involves nice and provable padding techniques, such as Shoup's optimal asymmetric encryption padding (OAEP). Another drawback to the book is its absence of information about hardware security. Securing hardware implementations is a difficult task, which involves knowledge of various threats, such as compromising emanations, demonstrated by the program Tempest for Eliza, or several recent attacks on smartcards based on power analysis and fault attacks. Moreover, with the advent of wireless networks and the quickly developing usage of smartcards and radio frequency identification (RFID), these topics present a direct interest for most of us. In conclusion, I do not believe that this book reaches its goal of presenting a general introduction to computer security to a broad audience. Still, I would recommend it to students and researchers in cryptography who are seeking an overview of software security. However, this audience might as well read the author's other book [1]. Online Computing Reviews Service