by Patrick McLaughlin
Two of the articles in this month’s issue, beginning on pages 9 and 17, discuss the progress on the Institute of Electrical and Electronics Engineers (IEEE; www.ieee.org) 802.3 group’s efforts to specify a 100-Gbit/sec Ethernet protocol. Both articles also discuss (and Ed Cady’s article goes into particular detail about) the possibility of a 40-Gbit/sec standard being produced alongside the 100-Gig spec.
As this issue of the magazine was going to press, word came from the IEEE’s late-July meeting that, in fact, the 40-Gig standard got the go ahead and will follow the same path to standardization as the 100-Gig standard.
As you read the articles that address 100 and 40 Gig, please remember they were written-and, in fact, were produced and headed off to print-well before the IEEE’s late July meeting. They were done and gone when we got word about the thumbs-up vote for 40 Gig. So, the “potential” 40 Gig specifications you’ll read about in those articles will come to fruition.
My thanks go to the two authors, Andrew Oliviero of OFS and Ed Cady of Meritec, for their thorough reporting on the matter. And in particular, I express my gratitude to Oliviero for giving me the late word about the 40-Gig standard, so that I could, at the very least, get the news to you on this page.
Coincidentally, over the past month, I have become aware of activities in two other standards-making bodies that I think merit some discussion. You can read more about one of them beginning on page 31 of this issue, so here, I’ll just briefly tell you that the InterNational Committee for Information Technology Standards (INCITS; www.incits.org) is about to release a set of specifications for the transmission of Fibre Channel signals over Category 5e, 6, and 6A twisted-pair copper cabling, in a project called FC-BaseT.
Claudio DeSanti, who chairs the INCITS T11 Technical Committee, explained to me, “Most Fibre Channel physical layer modules are packed in the SFP form factor. The FC-BaseT project started with an investigation aimed at verifying if a 1000Base-T PHY could be used to carry Fibre Channel at 1-Gbit/sec speed, and was triggered by the appearance of 1000Base-T SFP modules.”
After discovering some network-level limitations that prohibit a 1000Base-T PHY from carrying Fibre Channel, “A new protocol definition was needed,” DeSanti continued. “The guiding principle of the standard development was to re-use as much as possible the 1000Base-T designs, extending them to run up to the 4-Gbit/sec speed, in order to make possible new implementations to be based on existing designs. Therefore, there has been a strong collaboration with IEEE 802.3, and some liaison also with TR-42.”
Finally, there will be a new version of the TIA-606 standard for the administration of telecommunications infrastructures. Currently in its first revision, 606-A, the specification set is commonly referred to as the “labeling standard.” Earlier this year, I believed the TIA TR-42.6 group simply would affirm the 606-A standard, which is one of three options (revising and rescinding being the others) available to them once a standard is five years old. Now, however, it looks like the group will revise the standard in a process that ultimately will produce TIA-606-B.
At least part of the impetus for revising is a current project by the group to produce an addendum to 606-A dealing with data centers. The addendum is meant to reconcile 606-A with the TIA’s 942 data center standard; 606-A did not consider data centers and 942 did not consider administration. The two concepts will get together in Addendum 1 to 606-A. Furthermore, the TR-42.6 group will move ahead with work on 606-B. As the group’s recent meeting minutes state, “The changes will include, but will not be limited to, extending the concepts provided in TIA-606-A Addendum 1 into spaces other than computer rooms and equipment rooms.”
We’ll keep you posted.