|
|
|
The Bookworm
by Peter H. Salus
<[email protected]>
Peter H. Salus is a member of ACM, the Early English Text
Society, and the Trollope Society, and is a life member of the American
Oriental Society. He has held no regular job in the past lustrum. He
owns neither a dog nor a cat.
ATM
In 1991 I got a book on ATM by Haendel and Huber; 1994 brought me a revised and augmented edition; 1995 brought me a book on ATM by Uyless Black. The past few months brought me a volume by Kercheval, three by Black (one of them the second edition of the 1995 volume), and one by Giroux and Ganti. So here's an attempt at an overview. First of all, it's important to know that ATM was the creation of the telephone companies, intended to fit neatly into the ITU's broadband vision of the future. But the data-communication community took a look at that version and decided that they could do better. ITU-T constructed a model not unlike the famous ISO seven-layer cake. In this model, the ATM layer is immediately above the "physical layer." ATM is asynchronous because each station in the net can send or receive as many or as few cells as it wishes. (Cells are the fixed-size packets used in ATM. Thanks to the CCITT they comprise 53 bytes: 5 bytes for the header and 48 for the "payload.") When ATM came out, I hated the concept. After all, IP is connectionless; if a link goes down, it gets routed around. In ATM, if a link goes down, you're down. But further work has convinced me that ATM will be with us for a long time in some form. Black's 1995 volume was subtitled Foundation for Broadband Networks. The second edition is just a bit larger, waxing from 426 to 446 pages. Volume II is Signaling in Broadband Networks; volume III is Internetworking with ATM. They are all interesting books, though I thought volume III weaker than the other two. Structurally, they all carry introductory material, lists of abbreviations, and an excellent list of references. In between they are quite different from one another. Foundation takes us through an introduction to modern telecommunications to layered communications involving SONET and ATM. We then ease our way by means of ISDN-B into ATM, covering the AAL, switching operations, traffic management, connection control, and internetworking, before returning to SONET. Black then covers network management, the physical layer, and (finally) the business model. He admits that the current ATM market is soft, which few other authors are willing to actually state. With Black's Foundations volume read, you can turn to Vol. II, Signaling in Broadband Networks. Here Black details the evolution of the differences between signaling and transport network, the blurring of distinctions between them, and SS7. SS7 (Signaling System 7) is both powerful and flexible, though it does (like POTS) use an out-of-band system for signaling. While it's true that SS7 was intended for physical circuits, it can be modified for employment with virtual circuits too. Black goes on to describe ISDN-B and then moves at full speed into broadband technology. While he does a fine job of explaining a vast number of topics, he employs an even vaster number of abbreviations (largely three- and four-letter ones) in his presentation. The result (for me) was that I was continually flipping to the list of abbreviations to ascertain what was being discussed. Sentences like: "ATM uses small, fixed-length units called cells that are identified with VPIs and VCIs that are contained in the cell header" do not increase readability. Nor do chapter titles like "SAAL, SSCOP, and SSCF" (chapter 6). At some point, we've got to chew our way out of the alphabetic spaghetti. But despite this, I think that Black does an admirable job of presenting the mechanics of broadband signaling as well as presenting a very useful list of possible error messages. I had more problems with volume III, Internetworking with ATM. This was largely because of Black's idiosyncratic use of "internetworking": "Internetwork-ing is the sharing of computer resources by connecting the computers through a number of data communications networks." I prefer Comer's version: "Internetworking accommodates multiple, diverse underlying hardware technologies by providing a way to interconnect heterogeneous networks and a set of communication conventions" (Internetworking with TCP/IP, vol. 1, 3rd ed., 1995). Another problem I had with volume III arose as a result of receiving Giroux and Ganti, an excellent book on QoS under ATM. As I happen to believe that QoS may be the most important question in networking in general, I was much taken by Giroux and Ganti's presentation of ATM traffic management. As a comparison, Giroux and Ganti devote nearly four times more space to frame relay on ATM networks than does Black. Their exposition is also significantly less riddled by abbreviations and is quite clear and well written. In fact, Haendel, Huber, and Schroeder devote more space to frame relay than does Black. These books are very different from one another: they serve more or less the same audience, but in different ways. For example, those of us devoted to TCP/IP will find Kercheval indispensable. Those wanting an overview will want Haendel, Huber, and Schroeder, and Giroux and Ganti. And those wanting a detailed rundown will go for Black's volumes.
Networking
In fact, even a few pages on WDM would have been welcome in most of the volumes in this column on ATM and other aspects of networking. Reading the stuff in <http://www.atmdigest.com/WDM.htm> and in <http://www.oiforum.com/index.html> (Optical Internetworking Forum) makes you realize that this is the true technology of the future. Mark Sportack has written a really interesting "comprehensive introduction to routing concepts and protocols." I enjoyed it a good deal; and I think I learned a lot, too. Sportack calls routing "the most complicated function of a network," and I'd agree with him. But there's a gap in his presentation: not a word on queueing or queueing theory. There should be. Len Kleinrock's two volumes came out over 20 years ago; more recently, there's one by Gross and Harris (3rd ed., 1997) and a short volume by Kleinrock and Gail (1996). Tripod has done a really fine job on Cisco routers. I especially enjoyed his chapters on switched circuits, the Cisco router inventory, and troubleshooting. It's a well-written, useful book. Oppenheimer's volume is a truly worthwhile one. From "Identifying Your Customer's Needs and Goals," through "Logical Network Design" and "Physical Network Design," to the chapters on testing and optimizing, I found myself admiring her presentation and thoroughness. And that's more than enough words on ATM and networking for a month or two.
Uyless Black
Uyless Black
Uyless Black
Natalie Giroux & Sudhakar Ganti
Rainer Haendel, Manfred N. Huber, and Stefan Schroeder
Berry Kercheval
Kevin Downes et al.
Mark Sportack
Mark Tripod
Priscilla Oppenheimer
|
 Need help? Use our Contacts page.
Last changed: 27 Oct. 1999 mc |
|
Issue index