14 internautes sur 17 ont trouvé ce commentaire utile
R. D Johnson
- Publié sur Amazon.com
First, a comment about the update, then the comments about the book itself:
This follows a trend I've noticed in updated editions recently, in that it doesn't actually bill itself as one other than where it says '2nd Edition' on the front cover. There is no 'Preface to the 2nd Edition' or list of changes. If you had never seen the 1st Edition you would have no clue this book was an updated edition. Same octopus on the front cover of the class Cephalopoda, approximately the same number of pages, and a lot of the same content. However, the authors wife Joann Zimmerman is now listed as a co-author, and in the first edition he lived with her and their cat Sophie. He still lives with his wife and cat, but the cat's name is now Mona. (Rest in peace Sophie.)
The purpose, I suppose, is the same as Stroustrup noted in his latest Edition of the 'C++ Programming Language', when rather than highlight the old and the new, he desired instead a snapshot of the current state of the language. This book is similar with respect to Ethernet. It covers the same territory as the 1st edition, so many of the reviews of that book are still pertinent. Many of the sections are almost identical. However, entire chapters and sections of the old book, such as 'Model Configuration Guidelines for 10Mbps' and the entire Appendix B on crimping, wiring and installing thick and thin coaxial media are entirely missing, replaced with a small chapter on 10Mbps Ethernet. Given that coaxial Ethernet is totally obsolete, this is not surprising. The 1st edition had pictures of PCI Ethernet cards, also now gone in the 2nd edition. A small section on Twisted-pair Ethernet and Telephone Signals is still present, but almost entirely rewritten. The fiber optic section has been updated and expanded, dropping old connectors no longer seen and adding new ones. The 1st edition covered up to 1000Base; this 2nd edition now goes up to 100GBase in some detail. There's a tiny chapter on 400G, but it's not much more than a description of the workings of the study group as of 2013. In summary, some now-obsolete parts of the old book are gone or much reduced, a good chunk of the book (eg the 1000Base stuff) is almost identical but with updated editing and terminology, and there's quite a bit of new content covering the changes to Ethernet since the 1st edition's publication in 2000. It's all been mixed, merged, and moved around however, so like with the Stroustrup book it's best if you look at it afresh. On the other hand, if you're still crimping coax connectors you'd be better off picking up the 1st edition (or finding a new job at a modern facility).
If you understood and appreciated the 1st edition you'll enjoy the 2nd just as much, and you won't be wasting your time reading copious details about how to efficiently route half-duplex connections or searching in vain for PoE information (which is discussed in detail in this 2nd edition btw). However, the niche this book covers hasn't changed from the 1st edition. This is a book which describes Ethernet as a layer 1 and layer 2 component of the OSI network stack. Just like the 1st edition, there is little coverage of higher-level protocols such as TCP/IP, NetBIOS, or the like. Those higher-level protocols could be just as easily transmitted by carrier pigeon (OK slight exaggeration perhaps), and Ethernet is just another type of carrier pigeon. If you need TCP/IP information, look elsewhere. If you want to configure a Cisco router, again, look elsewhere. Likewise, this book is not the definitive guide for Ethernet chip designers--there's a few thousand detailed specification pages from the IEEE you can read for that. This book is intended primarily for software, hardware and network engineers and installation technicians who need a broad, detailed understanding of the physical Ethernet infrastructure as it applies to cabling, switching, routing, and configuration at the PHY level. At that it excels.
The one thing this 2nd edition is missing that I wish had been included is some information on IEEE 1588 AVB, now part of IEEE 801.2AS and friends, and the associated IEEE 1722 layer 2 transport protocol. These standards were ratified as of November 2011, but perhaps the authors decided not to cover them as few vendors provide AVB support in their switches at the current time.