There’s not much question that the recently approved IEEE 802.3an 10GBase-T specifications, combined with the still-in-progress Augmented Category 6 specifications from the TIA, collectively make up the biggest bandwagon to roll through the structured cabling industry in a long time.
If you plan to attend the BICSI Fall Conference this month, expect to see 10GBase-T/Augmented Category 6 everywhere you turn. As you go through this magazine, you’ll have to flip page after page before you find anything other than 10GBase-T as the topic of conversation. You can’t escape it, even if you try.
Far as I can tell, as bandwagons go, this one is long and wide. But I’m not so convinced that all its passengers are leaning forward in unison as it rolls down a steep hill at breakneck speed-at least as far as cabling infrastructure goes. Consider the following a fair warning: As you read through one page after another (after another after another) of information on 10GBase-T, don’t expect everybody to be “on the same page” with their messages. In other words, you won’t finish reading this issue, put it down, and say to yourself, “I can’t wait to implement 10-Gigabit Ethernet, and thank goodness there’s only one way to do it so I don’t have to stress over choosing the best one for my situation.”
Fortunately or unfortunately, there’s more than one way to deploy 10-GbE. In fact, there’s more than one way to deploy its twisted-pair flavor, 10GBase-T. As is often the case in such technological matters, there are scientific facts upon which everyone can agree, opinions that vary widely, and a whole lot of stew in the middle-opinions based on facts-that we could, and maybe will, debate forever.
First, the easy part. Everybody with whom I have spoken agrees that 10GBase-T is new ground in many respects. The protocol will transmit at up to 500 MHz, a full five times the maximum frequency of its predecessor, 1000Base-T. Everybody also agrees that, thanks to alien crosstalk, testing cabling systems for 10GBase-T capability will be a real bugaboo. They don’t all use the term “bugaboo,” but I’m choosing to use it here because I realize the contents of this publication are a) available on the Internet, and b) sometimes left in places such as rest room floors-both of which are accessible to children. And I wouldn’t want kids to read the words some people are using to characterize what it takes to conduct comprehensive testing for alien crosstalk.
That characteristic-alien crosstalk-appears to me to be the tipping point that, even if it doesn’t necessarily bring the bandwagon to a screeching halt, almost certainly divides the passengers concerning the direction in which they should go. You’ll see what I mean throughout this month’s issue. Our monthly columnist, Donna Ballast (page 10), concludes her in-depth research report on 10GBase-T with what she fully admits is her own opinion: Consider media other than UTP if you’re thinking beyond 10 gigabits. Also, network-system designer Donald Latham (page 100) gives us a history lesson, including some body blows to Category 6, which he hopes we’re not doomed to repeat.
My sense, based on nothing scientific, is that skepticism of Category 6A’s long-term value is a minority opinion. Each UTP cabling category that has been specified in lockstep with an IEEE Ethernet standard has succeeded both technically and commercially. Categories 3, 5, and 5e had long and useful lives in the cabling marketplace. Category 4 was a technical success because it supported 16-Mbit/sec Token Ring, but a commercial disaster because the more-capable Category 5 quickly followed it to market and Ethernet won the day over Token Ring.
Category 6 has been, from what I can tell, commercially successful. Technically, it does everything that Cat 5e could also do, and as Latham points out, the promise of supporting Gigabit Ethernet over two pairs was reneged when the IEEE dropped plans to create that protocol.
Augmented Category 6 will fall into the former group (Cats 3, 5, 5e) rather than the latter (Cats 4 and 6), I believe, thanks in large part to its direct tie-in to 10GBase-T. What will intrigue me over the next few years will be whether those who invest in 10GBase-T-capable cabling go with shielded or unshielded.
PATRICK McLAUGHLIN
Chief Editor
patrick@pennwell.com
