--By Arlyn S. Powell, Jr.
At the annual conference of the Building Industry Consulting Service International Inc. (Tampa, FL) held in Orlando, FL, this past January, the hottest question on the agenda was: Should there be a standard under development for Category 6 unshielded twisted-pair, or UTP, copper cable?
Paul Kreager, principal of Kreager Associates (Pullman, WA), who chaired a panel discussion on this controversial topic, summed up the situation: "Proponents of Category 6 say that Category 5 limits are already being pushed to the maximum, and we better start developing improved copper technology now. Proponents of fiber feel that any further standards work on copper distribution is an exercise in futility. Fiber is the ultimate transmission medium, and the faster end users move to it, the better off they will be."
The commercial building wiring committee TR-41.8.1 of the Telecommunications Industry Association, or TIA, (Arlington, VA) is responsible for the TIA-568 standard that defines the category system separating cabling and connectors into groups on the basis of performance. Paul Kish of Northern Telecom (St. Laurent, Quebec), who chairs the working group, stated that his group "is not currently working on a Category 6 specification. If there`s a need or an application for Category 6, our working group will entertain proposals. If there are no proposals, that will be the end of the issue."
Kish pointed out that proposals for Category 6 UTP are coming from European cable manufacturers and not the United States. "One of these proposals," he added, "calls for an extended frequency range to 300 megahertz, attenuation similar to that of Category 5 extended to 300 MHz, near-end crosstalk loss about 20 decibels better than Category 5`s extended, cable transfer impedance that is specified, and a cable core that is screened, with cable elements that may be individually screened as well."
Several technical issues would have to be addressed in a Category 6 specification, according to Kish. Manufacturers of extended-performance Category 5 cabling have focused on attenuation-to-crosstalk ratio, or ACR, to define the capabilities of their products, but with Category 6, says Kish, "ACR is not the whole story. External noise coupling is also a major issue and will call for consideration of balance between twisted pairs, screening effectiveness and grounding factors. Also, a frequency of 300 MHz corresponds to a wavelength of less than 1 meter; this means interaction between mismatched components will be critical, and factors such as impedance mismatch and resonance will have to be considered."
There are also commercial issues that would have to be considered. Category 6 systems are projected to cost 50% to 100% more than a Category 5 system. A nonstandard interface may also be needed because the 8-pin connector could prove to be inadequate for a higher performance level. Better-performing baluns may also be needed to permit link testing of a Category 6 system.
Kish concluded that, at present, transmission performance of a 300-MHz channel cannot be assured. "A Category 5 channel can support 155 megabits per second," he said, "and 622 Mbits/sec is feasible with encoding and by using all four pairs. But it is premature to specify higher-performing cable without also specifying the connecting hardware and channel performance."
The case for optical fiber was presented by Martyn Easton, the fiber-to-the-desk product manager at Siecor Corp. (Hickory, NC). Pointing out that Category 6 would be still another step in the copper-recabling treadmill, he said, "Copper cabling continues to reinvent itself with incremental performance gains, each time stretching the physical limits of the technology and further reducing the robustness of the solution.
"The copper camp is ramping up to meet multimedia bandwidth needs," Easton added. "This will create a need for recabling, as well as new termination, installation and testing procedures."
Fiber, on the other hand, can deliver 2.5 gigabits per second in premises applications, comfortably accommodating all of today`s networking needs. Attenuation is not dependent on transmission frequency, electromagnetic interference is not a problem, and installation, testing and certification are straightforward. "Fiber clearly offers the lowest long-term cost of ownership," he concluded, "and its performance is vastly superior to that of any grade of copper."
Acknowledging that the cost of fiber-optic network electronics was still high compared to copper, Easton called it a Catch 22 situation. "If there is no cable, there will be no electronics, and vice versa," he said. "Right now, there are no inherent barriers to low-cost, high-speed fiber electronics. It is simply an issue of sales volume being needed to bring the price down."
The most vocal champion of a Category 6 specification was Robert D. Love, a senior engineer at IBM`s Network Hardware Division (Research Triangle Park, NC). Love said that Category 5 specifications had been developed by and for cable manufacturers--and not for users. As a result, he stated, "key performance parameters such as far-end crosstalk, noise immunity and electromagnetic compatibility are either missing or inadequate. Category 5 performance is significantly limited by connectors and termination procedures, too.
"Applications are already pushing the limits of Category 5," Love added. "We have an opportunity to develop a Category 6 specification that meets all our critical needs." Among these are simple installation and termination and end-to-end testing.
He pointed out that proponents of fiber trumpet it as a one-time solution, but the emergence of singlemode and plastic fiber as possible replacements for multimode fiber makes that claim debatable. Also, claims that the cost of fiber installation is declining should not be allowed to mask the fact that the cost of network hardware and the range of selection are much better for copper than for fiber solutions.
"Category 6 development will take 3 to 5 years to do properly," Love concluded. "We have that time now to develop a cable specification to meet the needs of all parties."
After listening to the extensive discussion following the formal panel presentations, Kreager had several observations. "The most important point to make here," he said, "is that this is the first time that an audience has been formally introduced to what the standards community is doing. Category 6 is not even on our agenda; the impetus is coming from the vendor community."
Kreager saw as a compromise full development of both Category 5 and optical-fiber standards during the next few years. If a higher-performance alternative to Category 5 cabling is needed, IBM`s Type 1A cable already exists. "We`ll always have low-speed applications for copper, with fiber there for backbone and high-speed applications," he said.
There will be no competition from wireless technology, according to Kreager. There are dead spots, line-of-sight problems and interference to contend with within buildings. Also, the bandwidth is just not there.
"The standards community is committed to making Category 5 work," he concluded. "My interpretation of the sentiment expressed here today is that we not push for Category 6."