Examining shielded-cabling for ultra-high-speed transmission
Recent lab testing and other research keep shielded systems in the game for today and tomorrow.
The long-running story for shielded twisted-pair cabling systems, in the North American marketplace at least, has been that of a medium whose performance has never been questioned but whose implementation has never quite caught on for practical reasons. Perhaps the two most significant reasons are cost and an ease-of-use reputation that has been debated. Shielded twisted-pair cable in any construction—F/UTP with a single foil over the four twisted pairs, S/FTP with a braided shield over four wire pairs that themselves are foil-shielded, or any other shielded construction—is a more expensive product than unshielded twisted-pair cable of the same category. The cable type also can be more labor-intensive for installers to work with, and generally is not as flexible as unshielded twisted-pair (UTP) cable. Plus, historically, networking protocols that shielded cabling supported were also supported by UTP.
North America’s rejection of shielded cabling has not been matched in all parts of the world, as evidenced by the International Organization for Standardization’s (ISO; www.iso.org) development of Class F/Category 7 specifications as well as Class FA/Category 7A specifications while the Telecommunications Industry Association (TIA; www.tiaonline.org) has not progressed toward a Category 7 specification. ISO’s Class F/Category 7 specifications, finalized in 2002, define system performance to 600 MHz and its Class FA/Category 7A specifications define system performance to 1,000 MHz; both standards describe fully shielded cabling systems.
10GBase-T takes the stage
The notion that “anything shielded can do, unshielded can also do” was called into question during the early development stages of cabling systems meant to support 10GBase-T, the twisted-pair version of 10-Gigabit Ethernet. The most difficult electrical performance characteristic associated with 10GBase-T was and remains alien crosstalk—noise from what have become known as “disturber” cables making its way onto what have become known as “victim” cables. As twisted-pair cable manufacturers put in significant research-and-development efforts to produce unshielded cables that could handle alien crosstalk well enough to successfully carry 10GBase-T traffic, many North American users began considering using shielded cabling systems for the first time in a long time, or the first time ever.
Although the marketplace is now populated with many options of UTP cables that meet the final Category 6A performance specifications, considerations over whether to use unshielded or shielded systems to support 10GBase-T remain.
Recently the AMP Netconnect division of Tyco Electronics (www.ampnetconnect.com), which has been a long-time advocate of using shielded systems for 10GBase-T transmission, publicized results of testing conducted at an independent lab. Those tests compared Class EA (Category 6A) systems of unshielded and shielded constructions in terms of their performance when carrying live 10GBase-T traffic. According to Tyco, the testing included two UTP systems, one F/UTP system, and two S/FTP systems.
It would surprise no one that an advocate of shielded systems would produce test results showing that shielded outperformed unshielded. And that’s exactly what the results indicate.
Conducted by third-party test laboratory GHMT AG (www.ghmt.de) located in Germany, the study compared the five different cabling systems, which were from five different manufacturers. Key findings that Tyco mentioned in its announcement of the study included that shielded systems offer high coupling attenuation and therefore high levels of alien crosstalk and electromagnetic compatibility performance.
By carrying out the study and publicizing these and other results, Tyco appears to be making the point that on an application for which there is a standardized UTP solution, shielded cabling is a superior choice from a performance standpoint. Tyco established the Web site utp-vs-stp.com, at which it provides layers of detail on the study, some test results, and conclusions that Tyco has drawn based on the data.
Once the study results reached social media networks and specifically cabling-industry groups within those networks, the fur began flying and as of this writing it still is. Much of the banter, not surprisingly, has been between representatives of Tyco and representatives of other cabling-system manufacturers. The debate has centered around some now-familiar topics related to 10GBase-T and shielded cabling, including alien crosstalk and earthing/bonding of shielded systems. Some of it has also focused on the study itself and the fact that just two UTP systems were tested, and whether or not it is valid to hold those two systems up as representative of all Category 6A UTP cabling systems. Critics of the Tyco study are also saying that this lone test has produced worthwhile data, but significantly more testing must be done for results to be considered repeatable.
40 and 100-Gbit
Meanwhile, some proponents of shielded cabling have begun to turn their attention toward the next generation of Ethernet-based transmission: 40- and 100-Gbit Ethernet.
Currently the Institute of Electrical and Electronics Engineers’ IEEE P802.3ba Task Force (www.ieee802.org/3/ba/) is developing Ethernet protocols with operating speeds of 40 and 100 Gbits/sec, and the group expects to complete the project and obtain final approval in June 2010. On November 20, 2009, the IEEE 802 Executive Committee approved forwarding the draft of the standard for Sponsor Ballot, the final of the two balloting stages in the approval process.
The only copper-cabling type recognized in the 802.3ba project is twinaxial, Infiniband-style cabling. No twisted-pair media, shielded or unshielded, is accounted for in the draft standard. The twinaxial cabling configurations allow only for short-reach transmission. The 40- and 100-Gbit specification calls for longer-distance transmission over fiber-optic cabling exclusively.
Last summer, a group of engineering professionals from the cabling and networking industries convened with researchers at Penn State University in a two-day workshop entitled “Greater than 10 Gbps Copper Ethernet Workshop” (cictr.ee.psu.edu/WORKSHOP/). The workshop was sponsored by Nexans/Berk-Tek research center in Pennsylvania, the Nexans Data Communications Competence Center (DCCC). “The research conducted by Penn State covers the theoretical feasibility of going beyond 10G, up to 100G over copper twisted-pair,” according to Dave Hess, technical manager in research and development at Nexans DCCC (www.nexans.com) and a workshop presenter.
“The focus of the workshop was to investigate general technical feasibility, cabling requirements, and market drivers in support of Ethernet LAN speeds greater than 10 Gbits/sec,” states Valerie Maguire, global sales engineer with Siemon (www.siemon.com), who also presented at the workshop.
The general sense from workshop participants was that the prospect of transmitting 40- and 100-Gbit Ethernet traffic over twisted-pair cabling for some distance is technically feasible based on theory and modeling, but the twisted-pair cabling would have to be fully shielded.
One presentation addressed the standardization process within IEEE 802.3. According to at least one workshop participant, however, proponents with an ambition to get shielded twisted-pair cabling specified into an IEEE 40- and 100-Gbit Ethernet standard may face a tough challenge. Geoff Thompson, a consultant with deep roots in the IEEE 802.3 heritage, explains that such proposals to IEEE must meet five criteria to be accepted. “Three of the five are traditionally the most difficult to meet and are interconnected,” Thompson says. “They are: broad market potential, technical feasibility, and economic feasibility.”
While those criteria may not appear to be instant roadblocks for providers of shielded cabling systems, IEEE 802.3-based standards cover hardware-system signaling, not cabling systems. When viewed from a cabling perspective, the technical and particularly the economic feasibility may be simple when compared to the challenges facing manufacturers of semiconductors. The technical and financial investments these companies would have to make may be deal breakers for the prospect of 40- and 100-Gbit Ethernet over twisted-pair cabling, particularly considering the smaller size of the market at the core of the network.
More market and technical research will be conducted before a decision is made concerning whether or not shielded twisted-pair cabling will ever support standardized transmission speeds of 40- and 100-Gbit/sec. Today, the medium remains a part of the discussion as 10GBase-T network switches begin hitting the market.
PATRICK McLAUGHLIN is chief editor of Cabling installation & Maintenance