I was reminded this week of how much perception is driven by perspective. In this case, it was because of our advocacy of FCoE. I was exchanging messages with one individual who interpreted this as an attempt to undermine Fibre Channel (FC) and send it to an early grave. At the same time I was exchanging messages with someone else who felt we should not be wasting out time on FC and be spending more time and effort on IP-based storage. Needless to say, I found the contradiction entertaining, but I thought it might be worthwhile exploring these sentiments a bit.
“Doesn’t Cisco want to get rid of Fibre Channel?”
This one is easy–nothing could be further from the truth. We are committed to FC for the long haul because, simply, our customers are committed to FC. At the end of the day, in the enterprise, FC is still the standard against which other solutions will be judged for performance and availability. Even if customers make the decision to adopt IP-based storage, there is going to be a huge amount of data thats going to stay in the FC domain. It may stay put or be migrated slowly as part of normal refresh, but the end result is that FC is not going away anytime soon. From our perspective, we will continue to invest in FC as long as our customers tell us its important. Lest you doubt that, look at the updates to our Cisco MDS family over the last year and also remember that we still sell gear with Token Ring interfaces.
“Why spend time on Fibre Channel protocols?”
This is a fine question. To paraphrase bank robber Willie Sutton, we’re investing the time in FCoE because that’s where the data is. One of our primary data center design tenets is a unified fabric at the access layer for its TCO and functional benefits. We are agnostic about how you do that, whether its via IP-based storage or FCoE. From a practical perspective, as noted above, for most enterprise customers, their data is sitting in an FC domain, so any convergence strategy needs to take that into account. And while the storage folks may not care what we are doing at the server access layer, they are certainly not looking for their lives to be made any more complicated. Hence, we have FCoE.
At the end of the day, storage strategy shouldn’t be technology-dependent. The next-gen data center is going to need to support the ability of apps to grab data wherever it happens to be sitting: on IP-based storage, FC-based storage, or in a cloud somewhere, which is what we are ultimately helping our customers prepare for.