While a TIA subcommittee has made some decisions about Category 8, many more remain to be made as the specifications take shape over the next two years.
by Patrick McLaughlin
The topic of a Category 8 cabling system, engineered specifically to support 40-Gbit/sec data transmission over twisted pairs, has advanced from being a concept to being a work-in-progress. And while the topic has gained much attention and spurred discussion, the fact is that anything resembling a solid set of Category 8 specifications remains on the distant horizon. Some estimates project the timeframe in which the Telecommunications Industry Association (TIA; www.tiaonline.org) TR-42 Telecommunications Cabling Systems Engineering Committee will finalize Category 8 specifications to be late 2015. By virtually any measure, fall 2013 is very early in the process for the TIA and other standards-development organizations--the International Organization for Standardization/International Electrotechnical Commission (ISO/IEC) in particular--as they conduct the work necessary to specify a twisted-pair cabling system capable of supporting 40 Gbits/sec.
This article will address what is "known" and what remains "unknown" about the development of Category 8 specifications. It will focus primarily on the TIA's development of Category 8. However because those efforts are related to cabling-standard efforts within ISO/IEC, and to the Institute of Electrical and Electronics Engineers' (IEEE) development of 40GBase-T specifications, the article also will discuss those organizations and some of their current activities.
40G the driver
Driving the development of Category 8 cabling specifications is the work of the IEEE's 802.3bq Task Force, which is on the path to specifying the transmission of data at 40-Gbit/sec over twisted-pair cabling--40GBase-T. As Siemon Company (www.siemon.com) stated in one of the posts on its Standards Informant blog, the 802.3bq Task Force was "formed in March of 2013. The target publication date, as noted on the group's Project Authorization Request, is February of 2016. The Task Force has active liaisons with the TIA and ISO/IEC to ensure that cabling requirements under development will support the application."
Within the TIA, work on Category 8 is taking place within subcommittee TR-42.7 Telecommunications Copper Cabling Systems. As we reported earlier this year (see "Category 8 dominates cabling standard developments," May 2013), TR-42.7's decision to move forward with the nomenclature Category 8 raised the possibility of market confusion and a potential for challenging coordination with the ISO/IEC JTC1 SC25 WG3 (Joint Technical Committee 1, Sub Committee 25, Working Group 3), which also is developing a set of twisted-pair cabling specifications to support 40GBase-T. Since that article was published, the work of the two groups has come closer together rather than diverging. Following the TR-42.7's mid-June meeting, Siemon Company stated via the Standards Informant that 42.7 "accepted the concept of adding ISO/IEC Class II cabling performance criteria to its pending ANSI/TIA-568-C.2-1 Category 8 project. The Subcommittee also agreed to create a task group … to work on developing this criteria."
Answering the question, just what is Class II cabling, the Standards Informant explained, "Class II is the name of the new ISO/IEC grade of cabling that will be constructed from fully shielded ISO/IEC Category 8.2 cords, cables and connecting hardware. Both Class II and Category 8.2 specifications are targeted to support the 40GBase-T application over a distance of at least 30 meters and are under development by the ISO/IEC JTC1 SC25 WG3 Working Group. Category 8.2 components will be an extension and superset of existing Category 7A components."
Concerning the implications of the TIA's including Class II performance in its Category 8 standard, the Standards Informant added, "By accepting the concept … and creating a task group to work on these limits, TIA is demonstrating that North American standards development organizations are ready to embrace fully shielded cabling systems. This is a strong and positive step toward global harmonization of the full suite of available IT network structured cabling solutions."
Prior to the TIA's addition of Class II performance criteria, one of the realities of its earliest proposed Category 8 specifications was that for some electrical performance parameters at some frequencies, the Category 8 performance requirements either were no-more-stringent or less-stringent than the ISO/IEC's Category 7A/Level FA specifications. Though the TIA never drafted a set of Category 7 or Category 7A specifications, this quirk in the originally proposed Category 8 performance limits raised the ire of some because it would have meant TIA's Category 8 would not be a performance superset of ISO/IEC's Category 7A/Class FA. Those concerns were relieved by the TIA's action on Class II.
Into the unknown
Generally speaking, the TIA's decision to develop Class II performance specifications is the threshold of what is "known" about how its Category 8 standard will take shape. Sterling Vaden, who most recently chaired TR-42.7 and as of October 2013 serves as the subcommittee's vice-chair, is vice president of research and engineering with Optical Cable Corporation (OCC; www.occfiber.com). When I asked Vaden about the alignment of Category 8 with Class II specifications, he reminded me, "It amounts to potentially adding specifications for Class II, not changing the fundamental Category 8 requirements. There may be proposals to change some of the Category 8 requirements as we go along. It is a flexible, living process. For example, there are proposals to adjust the channel insertion loss, but they are not based upon ISO/IEC harmonization."
Other "knowns" relate to the established ISO/IEC Category 7A/Class FA specifications, but exactly how those specifications will mold what ultimately becomes TIA Category 8 remains to be seen. Understanding the significant caveat that the specifications are essentially two years away from completion, one of the very early "unknowns," or sources of speculation, is the Category 8 interface. Again, ISO/IEC's Class F/FA is the established reference point.
Another statement made in Siemon's Standards Informant is as follows: "The connecting hardware interface to support this new level of cabling [Class II] has not yet been specified by TIA. However, it is the opinion of the cabling experts at Siemon that the 8-position RJ-45 modular interface does not exhibit sufficient performance margin to support new requirements based on ISO/IEC Class II cabling." It then references Siemon's own TERA connector as an "interface that is already standardized by ISO/IEC." TERA is one of three interfaces standardized in ISO/IEC's Category 7A spec. The Standards Informant further states, "Fully shielded balanced twisted-pair connecting hardware characterized to 2 GHz [including TERA] would be ideal to support this new TIA level of cabling."
A different technical viewpoint has been published by CommScope (www.commscope.com) in its white paper titled "Category 8 Cabling Standards Update – TIA TR 42.7 Study Group for Class II Limits." In that paper, CommScope explains, "The key difference between Class I and Class II cabling is the fact that Class II allows three different styles of connectors that are not compatible with one another, or with the RJ-45 connector. This presents a dilemma to the customer who may not have detailed information on the advantages of each connector family. And, since the overwhelming majority of equipment ports, or MDIs, continue to use RJ-45, Class II cabling will require hybrid equipment cords. Class I, on the other hand, uses a fully backward-compatible RJ-45-type interface with vastly improved performance relative to Category 6A connectors and is specified up to 2000 MHz."
The paper also points out that ISO TR 11801-99-1 Class II specifications "do not cover connector types and, by default, refer to ISO 11801 for standard interface connectors. This means that three different types of connectors allowed in ISO 11801 and ISO 24764 data center cabling for Category 7A--including the GG45, ARJ45 and the TERA connectors--are allowed in Class II. If the GG45 connector is used, Class II cabling would be backward compatible to other categories if the GG45 jack is mated to an RJ-45 plug. However, the GG45 plug is not backward compatible to the RJ-45 jack; so, existing cabling with RJ-45 interfaces cannot be used with GG45 cords. In addition, the TERA and ARJ45 connectors allowed in the Class II standards will require the use of adapter cords to interface to existing cabling and equipment."
This terminology--TERA, GG45, ARJ45--as well as the interfaces themselves, may be largely or entirely new to many cabling professionals in North America. The Category 7A-capable technology has been largely moot in North America because the ISO/IEC 11801 standard is applied primarily outside of the continent. However, the TIA's development of Category 8 might bring this technology, and terminology, to the forefront. Might.
I asked Vaden about the dynamics of these three interfaces by way of a question that was really more of a statement. I asked (said), "Category 7A components do not meet the definition of interoperable. A user cannot take just any Category 7A plug and mate it with any Category 7A jack. The interoperability of plug and jack must be considered; it is not guaranteed." Then an actual question worked its way into the conversation when I asked, "Is TR-42.7 considering, or is it already pursuing, adopting multiple interfaces for Category 8?"
He responded, with more patience than was merited and with the experience of many standard-development projects behind him, "TIA could reference the IEC connector specifications, but they only go to 1 GHz. IEC has not begun to specify the connector performance. Yes, there are compatibility problems. It depends which way you look--forward or backward. The ARJ45 is the least-backward-compatible of the RJ-45 lookalikes, but the GG45 is not without problems either. Since the ISO Class II specs are for channel only, TIA likely will have to do the developmental research to add formal plug performance and mated connector performance criteria to the Category 8 document, as was done for Category 6A connectors. It is a daunting task, as there is no precedent upon which to build. … The connector issue will be affected greatly by the MDI [medium dependent interface, commonly referred to as equipment ports] performance, which is outside of TIA scope. The MDI performance will likely dominate all of the channel impairments aside from insertion loss. The box vendors will have to decide what type of connector to use. There are a lot of variables, and many of them do not involve the connector performance to a great extent."
Miles to go
While the TIA's Category 8 standard literally is years away from being finalized, glimpses of the technical debates that take place within TR-42.7 and other groups are visible to the general marketplace. Discussions about Category 7A interfaces, Class II performance levels, and interoperability/backward-compatibility are happening now in multiple forums and publications.
One of Sterling Vaden's final comments in my interview with him was, "As in previous generations, it may pay the user to wait until the dust settles a bit before making investments." At the same time, keeping an eye on that dust storm may be prudent in order to have a better idea of when the settling will take place.
Patrick McLaughlin is our chief editor.
Archived CIM Issues