Data Centre Ethernet Standards Fight Brewing ? Brocade Says FCoE in 2010.
7 July, 2008 by Greg Ferro Print Posting
I note this quote from a recent article on The Register where Brocade talks about their future strategy.
CEE - Cisco DCE is proprietary
“CEE is supported by 16 vendors committed to an open standard via the IEEE and IETF. It dismisses DCE (Data Center Ethernet) as a Cisco-trademarked term and expects it to be a proprietary implementation, not backwards-compatible either, meaning a fork-lift upgrade.
The CEE timetable is: 2010 - pre-standard CEE from start-ups; 2009 - standards completed and Brocade products introduced; 2010/11 - initial custmer trials and deployments”
Cisco recent Data Centre announcements all rely heavily on Data Centre Ethernet especially their FCoE strategy. If Cisco cannot bring the other vendors into agreement, then Fibrechannel over Ethernet is going to take a lot longer to happen, if ever.
Brocade Fibrechannel over Ethernet starting 2010
There will be FCOE concept validation in 2008/9, greenfield server-edge deployments starting in 2010, second generation FCOE products in 2011/12 and new FCOE storage perhaps appearing in 2012/13.
Holy cow, Cisco has been telling me that FCoE is here today, but Brocade doesn’t seem to think so.
Wrap Up
I was actually starting to believe the Cisco marketing around FCoE, but if the storage market is not getting behind it then it is starting to look in doubtful. I have previously pointed out the FCoE won’t happen this year, or the next, but to see Cisco’s competitors
I have seen a number of articles about Infiniband and iSCSI - is this the storage marketplace moving against Cisco ?
Ah well, I am sure it will all come out in the wash.
References
Block and Files - Original Article - check out the PDF download
Where Are All the Features for Nexus ? Or Is It Just Me ?
Is the Cisco Nexus 7000 Needed Today - or Tomorrow ?
Greg, FCoE is already available in limited footprint: the Cisco Nexus 5k supporting 10GE NICs from several vendors can connect FCoE to an existing FC SAN.
The really interesting thing to watch will be how this footprint grows as datacenter ethernet technologies are developed and implemented. The work being done in the IEEE and IETF is being led by folks from Cisco, Brocade, as well as others (don’t dismiss the start-ups!), and they all seem to be taking slightly different approaches to making their products available to the market.
For some of them, pre-standard or proprietary features will be necessary to their strategy. But over time I believe we’ll see the same pattern that we’ve seen many times before: they’ll all converge on a standard after it settles down (and they’ve claimed their piece of the pre-standard market).
Sure, you can clearly see that, but it isn’t here now. It won’t be here this year, or next year. If you believe the hype, or are deceived by it, you might buy first generation equipment that will need replacing (like the Catalyst 6000).
The marketing message is ‘FCoE is here and you should be buying’, Cisco sales reps are certainly be led to believe that they have something to sell. Cisco certainly wants the product put onto customer roadmaps, and is making a big push into getting their message across.
But fundamentals are that
1) DCE is mandatory for FCoE
2) DCE will make iSCSI work very nicely indeed
3) iSCSI will be more appealing to the market than FCoE and thus likely to overtake FCoE.
4) DCE is barely started on the standards process, and it only takes one major player to derail and delay it.
5) Infiniband already outperforms FCoE and offers a better future than iSCSI.
It hard to see FCoE being successful.
Greg:
I guess its helpful to do a couple of things. First, we need to separate conjecture from fact and second, we need to separate a given vendor’s product roadmap from the pace of innovation of the industry as a whole.
Let’s start with the assertion that DCE is proprietary, etc and will cause all manner of evil. This would fall under the category of conjecture, and wishful conjecture at that. Both “CEE” and “DCE” are trademarked terms, just different trademark holders. The actual IEEE term is Data Center Bridging (DCB - http://www.ieee802.org/1/pages/dcbridges.html), and, as with other aspects of the Ethernet standards, Cisco is deeply involved in the standards process behind DCB. The bottom line is that Cisco is focusing its efforts on delivering a standards-based solution. End of story.
As far as the timeframe? Well, pre-standard solutions are available from “startups” like Cisco and Intel today. Cisco is shipping the Nexus 5000 and Intel is shipping 10GbE FCoE adaptors today. Emulex and Qlogic have already announced their Converged Network Adaptors (CNA) and expect them to ship by late summer. Finally, Paul Brown, VP of Integration and Interoperability for EMC has stated “Today EMC is working with Cisco to qualify the Cisco Nexus 5000 in our E-Lab in order to provide customers with a complete unified fabric solution, targeting customer deployments in 2H CY2008.” For more details on vendor interoperability testing, check out Deepak Munjal’s post: http://blogs.cisco.com/datacenter/comments/setting_the_record_straight/
In my experience, we are seeking traction on a couple of fronts. The first areas is greenfield deployments where customers are looking to build-out their next-gen infrastructure. However, we are also seeing significant interest for brownfield scenarios where unified fabric will help customers with sprawl, power, cooling, TCO, etc. Remember, one of the benefits of the Nexus 5000 is that it can be inserted in a fairly transparent manner so the hurdles for implementation are quite low and the benefits quite accessible.
With regards to the fundamentals listed in your response, I would respectfully disagree. I think the whole iSCSI vs. FCoE debate is a red herring. While I think the market will eventually buy-off on a unified fabric architecture, I don’t think the entire storage market, across all market segments, is so homogenous that it will converge on a single technology. iSCSI, FCoE and IB will all have their adherents, so I think that it is difficult to say one protocol is better or more appealing than another–it all depends on technical, financial and business context. Heck, I am betting that a) plain old FC will still have a healthy presence 10 years from now, and b) many customers will have a mix of FCoE and iSCSI running across their Ethernet networks.
I guess we could chat a bit about the long-term viability of FCoE, but I think this response has gotten a bit long as it is.
Instead, perhaps you can check out this post on Chuck Hollis’ blog: http://chucksblog.typepad.com/chucks_blog/2008/04/fcoe-gets-buzz.html