Over the last few years the mantra from the computer industry seems to have been virtualize, virtualize, and virtualize. You can’t go to any conference without hearing about how the new widget helps you virtualize better. Whether it be a new server, a new switch, a new card, a new processor, or a new chip of some kind. Each of them will help you virtualize better.
So over the last two to three years most IT shops have enacted virtualization of one form or another. This year, you can expect the new buzzword of FCoE to be tossed around and it’s the new consolidation point. That’s where I came up with the name for this blog of ‘Down to the Wire’. Because now, consolidation is down to the wire that provides the critical data for your virtualized servers.
That being said, it’s still the fourth month of 2009. I’m enjoying my third full month with Emulex as part of the IBM OEM account team. We’ve done some education events here in the United States to update the technical community at IBM with regards to the 8GB roll out and how things look moving forward to the world of converged networking.
I’m a bit of a pragmatist when it comes to computer technology. While technology can change in an instant month after month it does take some time before you see solid adoption of a technology. Just look at virtualization and VMware as an example. I was doing road shows showing the awesomeness of VMware in 2001 on IBM’s eServer xSeries 440 platform in the trunk of my car, convinced this was the best thing out there to help with the proliferation of servers. It took a couple of years, but now virtualization is one of the top uses on servers compared to the beginning of the decade.
So why the short trip down memory lane? I’m convinced FCoE is a lot like virtualization. It’s going to be something people kick the tires with, play around with, check out in their test & development environments in an effort to cut costs there. However if you look at the data center infrastructure. 2010 might be the year you see deployments begin, proof of concepts. Why do I think this?
It’s a threefold issue.
The first is what I’ve called for years the ‘Layer 8 Networking problem’. The division of responsibility in most organizations is there is a SAN team and a Network team. 10G Ethernet, FCoE, iSCSI, and other technology straddling those two departments will create a nice battle royale at the manager layers to decide who does what.
The second issue is the infrastructure required for this. I haven’t met a lot of customers who are telling me that they want to rip out & replace both their SAN and Network infrastructure this week either. That, and the FCoE standard while in committee isn’t 100% solid. Of course the switch vendors out there still haven’t figured out that whole ‘open’ thing and are all trying to jockey & position themselves to tack on features that will make you ‘must have’ what they sell and who cares about anyone else. My favorite is the switch manufacturer that has bolted a switch chip to a PCIe card and told you they ‘work’ better because they’re the same chip. Never mind they’ve not written a device driver ever.
The final one is, the need to change. This is the one I think that will at least start the conversation. While virtualization has been an amazing tool to help manage server sprawl over the years, the sheer number of connections you want to an individual server have become an issue. The wad-o cables alone can cost thousands of dollars per server just for the rack. Don’t forget us poor saps who have to do that work too. With the delicate fibre cables and the four to six cat five cables in a regular configuration, or even that mass of cables on a blade chassis. Who wouldn’t like to get it down to a sleek svelte four or six little optical cables. Technologies like NPIV. (Which were developed by IBM and Emulex), and the new hardware offloading the Universal OneConnnect™ adapter promises to bring will really go a long way towards helping with this. However, I think 2009 isn’t the year that we’ll see broad spectrum adoption. It’s the year I think everyone will spend kicking the tires, and checking things out.
So I guess my thoughts for now, are go FC! It’s not dead yet!
[...] happy to see some offerings in this space as it starts the conversation. However like I have previously mentioned just because you brought the horse to water, you’re not going to force them to drink. As [...]
I think you are right, 2010 will still be a proof of concept/pilot year for FCOE. Typically things take longer than one imagines, as you point out 1) politics, and 2) installed base. In addition, the weak economy will probably delay some from investing in new technologies.
[...] my first post introducing my thoughts on where we stood with FCoE and it’s now been three months and there [...]