From the January, 2014 Issue of Cabling Installation & Maintenance Magazine
The twisted pairs that will enable the transmission of data at 40 Gbits/sec will be far from the only "balanced" aspect of delivering such high-data-rate communication.
By Patrick McLaughlin
With the calendar having turned to 2014, professionals in the cabling industry can--albeit with a fair amount of optimism--look forward to the possibility that next year will be the year in which the IEEE's 802.3bq 40GBase-T set of specifications, and the TIA's specifications for Category 8 cabling, are finalized. Both standards-development organizations are striving for late-2015 completion of their respective standards. As each of these organizations makes progress, more elements of each standard get solidified.
On October 31, 2013, Cabling Installation & Maintenance hosted a web seminar, during which Category 8 and other standard-development progress was discussed. Paul Vanderlaan, technical manager for standardization and technology with Berk-Tek (www.berktek.com) delivered a presentation during that seminar titled "Developments in Category 8." In addition to providing information on the goings-on within the TIA's TR-42.7 Telecommunications Copper Cabling Systems Subcommittee, Vanderlaan provided some perspective on the balancing act that is the creation of 40GBase-T and Category 8 cabling systems. Specifically, he explained that developing an application (40GBase-T) as well as a cabling system to support that application (Category 8) requires a sometimes-delicate balance of technical and market sensitivities. Three characteristics are particularly relevant: 1) reach, 2) bandwidth, 3) impairments.
Vanderlaan used 10GBase-T as an example of an application whose rollout was not as rapid or smooth as many would have preferred, using the three aforementioned characteristics as benchmarks. "With 10GBase-T, 100-meter reach was the holy grail," he commented. Previous-generation Base-T applications were specified to 100 meters, and with 10GBase-T being aimed at the enterprise as well as the data center, the 100-meter distance was an assumed necessity. The amount of power required to transmit 10GBase-T to 100 meters became a stumbling block, and 10GBase-T gained a reputation for being a power-hungry application. The development of Energy Efficient Ethernet and more-recent technological improvements to chipsets brought down those power consumption levels--years after the application's initial development.
For 40GBase-T and Category 8, the IEEE has targeted a reach of 30 meters. As a practical matter, 40GBase-T will be a data center application, not an enterprise-style, to-the-desktop application. As such, a 30-meter channel will accommodate network architectures such as end-of-row (or, as it often is set up in reality, middle-of-row).
The decision of how much bandwidth to designate for an application requires balance too. Though the term "bandwidth" often is used to mean throughput capacity, its literal meaning is the amount of spectrum (i.e. width of band) equipment uses to transmit and receive signals. 10GBase-T's bandwidth is 500 MHz. Vanderlaan observed, "If we use too little bandwidth and try to squeeze information into a smaller band of frequencies, it increases part complexity and size, as well as power [consumption]. If we us too much bandwidth we can extend too far. Some might ask, ‘Why not use 4 GHz for 40GBase-T?'" The issue, Vanderlaan pointed out, is that electronic components must process across every frequency in the spectrum at a speed of 40 Gbits/sec. He made a comparison between what it takes for high-speed transmission and what it takes for high-speed driving. "Think of a Formula 1 engine," he said. "If the engine constantly runs at 16,000 RPMs, it is using a lot of energy." In the case of a Formula 1 vehicle, that energy is fuel and a result of such consumption would be too-frequent pit stops for refueling. In the case of 40GBase-T electronics, conserving energy can allow the application to avoid the deserved power-hungry label that held back its predecessor 10GBase-T.
The balanced approach being taken toward 40GBase-T development, Vanderlaan noted, is to be judicious when choosing an upper bandwidth. Referring to too-low-bandwidth and too-high-bandwidth possibilities, Vanderlaan noted, "Either side is limited by realistic implementation barriers." The agreed-upon bandwidth for 40GBase-T is 2 GHz--four times 10GBase-T's 500 MHz.
When discussing impairments, Vanderlaan pointed out, "Cables, connectivity, circuit boards all contribute to budgets" for impairments such as return loss, crosstalk and insertion loss. A significant balancing act to be achieved in this realm is to improve performance by minimizing impairments, but not doing so at the expense of market acceptance.
Once again using 10GBase-T as a benchmark, Vanderlaan pointed out that today it is well-defined and shipping in volume. Power levels have been brought down and the consuming market is satisfied with the application's robustness. An objective, then, by the IEEE and TIA groups establishing the 40GBase-T and Category 8 standards, respectively, is to use 10GBase-T as a foundation or a stepping stone upon which to build.
The work is a high priority for each group. During his presentation, Vanderlaan noted that the Category 8 specifications will be published as TIA-568-C.2-1--the first addendum to the 568-C.2 standard. The group is putting off its work on the whole-scale standard revision, TIA-568.2-D, until it completes 568-C.2-1/Category 8.
Patrick McLaughlin is our chief editor.
Archived CIM Issues