A few thoughts as I watch the season’s first snowstorm up here in New England and wonder if I should pack my long johns for the annual BICSI Conference in Orlando, FL this January …
The IEEE has set its sights on 100-Gigabit Ethernet, as was announced in early December. There had been some debate as to whether the IEEE 802.3 Working Group would pursue 40-Gbit Ethernet next or 100-Gbit. It will be the latter. Furthermore, word is the group has agreed to support a reach of at least 100 meters on OM3-grade multimode fiber (or, as we like to call it, laser-optimized multimode). That’s 50-µm multimode fiber with an effective modal bandwidth of 2,000 MHz∙km at 850 nm. Obviously, much more will unfold in the months and years ahead, such as whether this speed will be achieved with serial or parallel optics. I plan to keep regular tabs on the situation, in part through the Ethernet Alliance’s homepage (www.ethernetalliance.org.) We’ll do our best to keep you informed through our pages, our Web site, and our e-newsletter as well.
The “wet link phenomena” remains an issue worth watching. An article on page 57 tells a lubricant supplier’s side of the story. It includes summarized results of some lab tests on cables in multiple environments. A couple of key points to remember as this situation continues to unfold: 1) the information in this month’s article will be presented, presumably in more significant detail than you’ll read in this issue, at a meeting of the TIA TR-42.7 Subcommittee; 2) the cables used in the testing were rated Category 6 and tested to Category 6 specifications, which are less stringent than Category 6A specs. It looks to me like we still have some time to go before “best practices” are established. Until then, I’d suggest the best practice you can employ is to ask the cable manufacturer which substances you can and cannot use to ease their product’s path through a conduit. Then document their answer.
Data centers are where 10-Gbit/sec Ethernet is seeing the most action these days, and very likely where 100-Gbit/sec Ethernet will see its first real deployment in however-many years it takes to develop. These hubs of processing, computing, and transmitting at breakneck speeds arguably have become an industry of their own over the past few years. They have their own bona fide conference and exhibition called Data Center World, which sometimes includes detailed information on cabling infrastructure. This month, we at CI&M magazine kick off a year-long series of coverage devoted to data centers. Each issue of this magazine in 2007 will include a feature-length article discussing data centers and, more to the point, cabling systems’ roles within them.
Shielded cable is making a comeback, at least in the realm of mindshare and perhaps also in market share. Last month, two authors from Tyco Electronics/AMP Netconnect flat out said shielded cable is a better option than unshielded for 10GBase-T (see “Why shielded twisted-pair matters for 10GBase-T,” December 2006, p. 13). At nearly the same time, the company announced a new training program, focused specifically on shielded twisted-pair cabling systems, for its certified design and installation contractors. Also recently, cable supplier Hitachi Cable Manchester reported that during the third quarter of 2006, sales of its shielded Category 5e and Category 6 products grew by more than 100% over the same quarter in 2005. Recently, somebody asked me if I believed we would see an unshielded Category 7 cable. My response was that we’ll barely have an unshielded Category 6A. In spite of all the efforts to defend against alien crosstalk with an unshielded Category 6A, it looks like some users are more inclined to deal with a foiled cable. As with everything else, we’ll see what really happens.
As always, I look forward to serving you in some way, shape, or form through the information in this publication and its Web site. For those of you whom I’ll see in Orlando, let’s hope the only things that are borderline frigid are the data centers.