Wishful Thinking: The FCoE Edition
Since there are folks out there that would have you believe we will see cold fusion before we see a Fibre Channel over Ethernet (FCoE) standard, I thought it would be a good time to provide an update, especially because the T11 working group tasked with developing the standard has been putting in some significant effort.To bring everyone up to speed, the FCoE standard is being developed by the T11 Technical Committee within INCITS. The T11 committee defines all the aspects of the Fibre Channel protocols. The work on FCoE started in April 2007, when T11 was led by Bob Snively (Brocade) as Chair and Claudio DeSanti (Cisco) as Vice-Chair. Sadly, earlier this year, Bob died and Claudio has been acting as interim chair. The actual FCoE development work is being carried out in the FC-BB-5 (Fibre Channel – Backbone – 5) working group, which is led by Claudio as Chair, and has Dave Peterson (Brocade) as Editor.The Working Group established a stable frame format in August 2007 and followed that up with an agreement on the addressing scheme in February 2008. Draft 1.03 of FC-BB-5 was sent to letter ballot in October 2008. The ballot closed on November 14th and the draft passed the ballot with a vote of 29 favoring, 4 opposing, 10 abstaining and 3 not voting. By the way, this vote represented a who’s who of vendors, including companies like Brocade, Cisco, Dell, EMC, Emulex, IBM, Intel, NetApp, QLogic, Sun et al. The vote yielded almost 700 comments, which needed to be resolved, although, to be fair, many of those comments were duplicates.The Working Group held some interim meetings in addition to the regular meetings and has managed to shrink the list of comments down to 8, primarily around reporting and management items. The current plan is to have these final 8 items closed out by the beginning of April, which would then set the stage for the final approval, possibly as early as this June. It is possible that Brocade will want to re-open discussion at the April meeting on some additional comments that are currently closed, but we do not expect any major turmoil.So, what is the practical impact of this. Well, with the frame format and addressing established early on, the hardware requirements are stable — for example, customers that have deployed the Nexus 5000 do not have to worry that their hardware will not support the final standard. As the rest of the standards are finalized, the changes will be implemented through software updates. Some of of the related protocols will be handed in a similar fashion. For instance, FCOE will require the implementation of FIP (FCoE Initialization Protocol). Cisco’s current implementation is not compliant with the standard, but will be brought into compliance with an OS update — again, without any hardware impact. DCBX will be handled in a similar manner. Cisco currently supports CIN-DCBX (Cisco-Intel-Nuova), will soon support CEE-DCBX (Converged Enhanced Ethernet), and will support IEEE-DCBX once the standard is finalized.Its also important to note, as Silvano Gai, who literally wrote the book on FCoE, pointed out, standards are organic things. If they are going to be useful and stay relevant, standards need to remain organic, so waiting for a standard to be “done” is kinda pointless.