potential of more breakthroughs in the future. The book by John Bingham is targeted at engineers who are involved in the design of DSL systems, and MCM.
I
I
I
NEW BOOKS and MULTIMEDIA
- > I
(Continuedfrom page 4)
‘
explaining the politics of I P address delegation, from its historical beginnings to theageofICA”andARIN(the Assigned Registry of Internet Numbers). In particular, it is explained how, along a delegatidn process that eventually reaches the user’s company, the viewpoint and objectives of ISPs is different from that of IRs (Intemet Registries).ISPs favor routability and maximization of customer base for t h e given allotment of I P addresses. The objectives of ISPs are different from those of the user’s company, in that users do not wish for frequent renumbering unless a business case is made to favor such changes. We arewarned that the task of renumbering is not a light one, especially when one considers the potential for loss of income while the system is in an inconsistent or unstable state. The casual reader is convinced t h a t t h e problem is indeed complex when more protocols and entities enter the picture, such as the Network Address Translation (NAT) ,the DNS (for example, the need for quick secondary DNS updates), the renumbering of routers (which can have impact even to t h e authentication of dialup users), etc. A chapter deals with each foreseeable complication. Recognizing that the best transition approach for renumbering depends on the scale of the site, Dutcher provides three case studies, for small, medium and large networks respectively. The concluding chapter looks forward to the future when the adoption of IPv6 will contribute towards solving most of the IP address management problems.
ADSl, VDSl and Multicarrier Modulation John A. C. Bingham, 2000, John Wiley and Sons, 0-471-29099-8, 289 pages, hardcover A part of the Digital Subscriber Loop (DSL) technology is based on advances in the area of Multicarrier Modulation (MCM), an area which still receives the attention of intense research with the potential of more breakthroughs in the future. The book by John Bingham is targeted at engineers who are involved in the design of DSL systems, and MCM modems in particular. The book provides an exciting personal view about the important future work in the area, motivated by the fact that its retiring author will not be working on them while feeling the importance for new research in certain areas. The chapters include sections describing the “Unfinished Business” of the corresponding topic. Clearly,,grad-
6
.
uate students and researchers will appreciate the guidance provided. A further feature to satisfy the intellectual appetite of readers are references to patents, t h a t may n o t b e known t o t h e wider audience. T h e curious r e a d e r finds abundant material that explains the particular complexities of high-speed data uses of the local loop, stemming from the impairments of the local loop (presented in c h a p t e r 3). Two c h a p t e r s introduce DSL systems in general and MCM respectively. The Discrete Fourie r Transform ( D F T ) based MCM is presented in the sixth chapter, followed by a chapter on alternative types of MCM (frequency domain spreading, filtering, time domain shaping and discrete wavelet multitone) discussing the tradeoffs of each technique. The implementation of DMT in ADSL is the topic of chapter 8. The coexistence of ADSL with other servicessuch as voice-band services, echo-canceled ISDN, and T D D ISDN are the subject of the next chapter. The bookconcludeswith achapteron the requirements and implementation of VDSL and a review chapter on future possible improvements which Bingham foresees to be seamlessly incorporated through the option exchange protocol defined in G.994. The appendices include MATLAB programs for xDSL analysis, a list of the relevant international organizations and standards and an appendix o n t h e efficient implementations of FFT engines.
Unix System Administration Handbook, 3rd Edition Evi Nemeth, Garth Snyder, Scott See bass and Trent R. Hein, 2001, Prentice Hall, 0-13-020601-6, 854 pages, softcover Large-scale installations of Unix systems are still some of the most powerful and intricately complex system to manage. The Unix System Administration Handbook is the rare exception of a book that devotes attention to all the issues a syste’m manager may encounter in a heterogeneous environment of Unix flavors. The third edition of the book reveals the gradual expansion of system administration tasks beyond backups and security, to the issues behind mainfaining web servers or managing the integration of Unix and MS Windows (for e x a m p l e , using S a m b a ) . T h e b o o k avoids trivializing the problems o n e should expect to encounter when managing a Unix site. Instead, it presents each problem at is representative level of complexity. T h e opinions of the authors are always stated clearly, discouraging
the reader to look into options that experience and t (me have rendered as inappropriate, 01 simply awaste of time. This candid approach to finding the best fix for a problem reveals a genuine appreciation for the fact that system managers are people who in t h e i r finite time have to come up with quick and flexible answers that make, almost, everybody happy. Considering, less than the best answex to a problem is simply not good for them. The obvious emphasis of the book i:j on identifying the best tools for the job, for each conceivable job and the balanced approach t o tradeoffs that may exist in coming up with solutions. The tl )pics are numerous to begin listing them but they contain an important and thorough description of network related configuration and troubleshooting, routing, electronic mail configuration, DNS, NFS etc. Whenvendors differ in their philosophy or just in the command line arguments, the book points the differencesout. The current editioncovers Red Hat Linux 6.2, Solaris 2.7, HPU X 11.00, a n d F r e e B S D 3.4. O t h e r systems ai e f r e q u e n t l y close t o t h e f o u r covered in t h e b o o k , with t h e n o t a b l e exception of A I X a n d SCO systems.
Java Server and Servlets, Building Portable Web Applications Peter Rossbi:ich and Hendrik Schreiber, 2000, Addison-Wesley, 0 2 0 7 -67497 2,429 pages, softcover
The universe of Web-based applications includes architectures that emphasize client-tide applications, “applets”, architectures that emphasize serverside applic,itions, “servlets”, and their combinatiions. T h e c o m b i n a t i o n of applets and servlets provides a wide spectrum of choices in deciding where and what computation is performed in a web application. O n e of the concerns in building servlets is the performance and the security that, ultimately, depends on the language used to implement the server-side computation. CGI scriptswere the first approach to server-side dynamic content generation. Proprietary server APIs developed as a result but the need for a flexible programming languagewith security features led to the natural use of Java and the definition of a servlet API between server and servlets. The book by Rossback and Schreiber covers the servlet AP[ 2.2. The philosophy of the b o o k is t h a t a useful servlet-based application invariably needs, a p a r t from the ability to use the servlet API and (Continued on page 8)
IEEE Network
JanuaryFebruary2001
