Are We Putting The Cart Before the FCoE Horse?

I am as big a fan of Fibre Channel over Ethernet (FCoE) as the next guy, but hang on guys, I think we are getting a little ahead of ourselves. It seems that lately we are already anointing FCoE king of the prom, yet it has barely picked out its tux. FCoE has great potential, but to advise IT managers that they need to move to it within the next two years is a bit premature.

First, FCoE and CEE standards are yet to be ratified. In theory, not much will change between now and ratification but they are NOT ratified standards. Most IT managers are not going to start deploying non-ratified equipment.

More important is understanding how the data center deals with change. Change in the data center comes at an expense measured in hard dollars to buy the technology and soft dollars to implement the technology. Adoption of new technology in mass happens when the pain associated with the current way of doing things is greater than those expenses. We saw that with server virtualization; it required money and effort to move to virtual servers but many customers feel that the effort is worth it.

I'm not convinced that we have reached that point with storage infrastructures. Do storage managers want more from their storage infrastructure and do they want to drive down costs while making it easier to manage? Sure. Who doesn't? I'm not convinced that FCoE early on will address many of those issues nor am I convinced that waiting will put you at a disadvantage.

What FCoE will address from the moment you install it is NIC and cable consolidation. That's it. Since most data centers will deploy FCoE slowly, very slowly, it will take some time for the positive effects of that cable consolidation to be measurable. The problem is the cost of the NIC and the cost of the CEE cable are more expensive, so any hard cost savings if not a total wash, will be minimized.

The payoff from FCoE may be years in the making, possibly five or six years into the future when most if not the entire data center is on a cost effective CEE infrastructure. That will also provide time for the tools to catch up with the environment so that you will truly have a single point of management and operation of the combined network.

The other issue that continues to bother me and I think threatens FCoE adoption is that the network and storage are two distinctly different groups in many organizations. I have been in several meetings discussing FCoE with a customer -- and we were introducing the network guys to the storage guys because they had never met before! Dealing with the politics of merging the groups together may take longer than merging the physical infrastructure.

With that, I don't think that you have to consider FCoE in this go around. If you want to get the move underway, that's fine -- but don't feel pressured to move to it today. If you need faster performance and want to stay with standard FC, then 8 Gbps is a proven tact to take and there is little chance of you being left out of the FCoE conversion movement when it happens.

As a result, pick vendors that are committed to both paths, traditional FC and FCoE. Make sure they have at least a 16-Gbps FC roadmap as well as a FCoE roadmap, make sure they have a solid mechanism for integrating mixed infrastructures. During times of change, you want all your options available -- you don't want to be forced into picking the winner.