After speaking with a number of storage and server vendors this week, it has become apparent to me that Emulex's much talked about "UCNA" is nowhere to be found. Some vendors who have standardized on Emulex's FC stack are growing increasingly concerned that their sole dependence on Emulex is causing them to be late to the game with FCoE. This means greater opportunities for Brocade and QLogic, who, as I mentioned in my previous blog, are well ahead of Emulex and its UCNA.
It was also brought to my attention recently that Emulex has a strange business arrangement with ServerEngines regarding the sales of the UCNA. Ultimately, Emulex will handle sales and support of the adapter version of UCNA, ServerEngines will handle all sales and support for the LOM version. Needless to say, this will create discomfort for customers as they have to deal with two different companies for the same converged networking technology--two support teams, two field teams, two service teams--which creates more stress for IT organizations and OEMs that will eventually deploy the technology. This business arrangement seems very strange.
At the same time, I understand that Emulex will not be participating at the FCIA FCoE demo at SNW this October, which just marks another instance in their continuing streak of missing FCoE events and plugfests. Keep in mind that Emulex and UCNA were both no shows at SNW Spring earlier this year; that Emulex had a picture of UCNA at EMC World in May, but no physical product; and that Emulex side-stepped on participating at the FCIA/EA Plugfest in May. The fact that they are missing from the November FCIA FCoE Plugfest at UNH confirms what I believe are ongoing delays with what appears to be an invisible piece of technology. Okay, is it really invisible? No, of course not...it's just very late.
And when it does finally surface, UCNA will be a difficult to work with. We understand that UCNA can't run Emulex's existing FC HBA drivers on a ServerEngines chip and without critical backwards compatibility it's no different than an Ethernet vendor trying to build a FC stack. Bottom Line: Without some very much needed clarification and reassurance to I.T. customers and OEM vendor customers, they are starting to believe UCNA may be so late that they will have to make other choices rather than risk infection of arrays with new and as yet field unproven technology.
Have U seen a UCNA? If you have, let me know. Take a picture of it to be proof positive, and let me know in what environment(s) you have tested.Tom Trainer is founder and president of analyst firm Analytico. Prior to founding Analytico, Trainer was Principal Storage Product Marketing Manager at Red Hat, and Director of Marketing at Gluster prior to its acquisition by Red Hat. Tom has worked as managing senior partner ... View Full Bio