In the news this week are interesting revenue reports that merit some attention, and coincide with a subject that in my opinion, needs a strong revisit. That subject is Cisco UCS, and how UCS stands poised to influence the infrastructure. (So Dave Simpson, despite what you said
, I don't think the Brocade earnings are boring after all!)
Before I say more, let's pause for a second, as I'd like to clarify I'm no UCS fanboy, even though it may sound like I am one in this post. I've long looked at Cisco's move into the server market with some ambivalence. You see, I'm not a plain vanilla kind of guy, and Cisco is rapidly turning some parts of the infrastructure (the ones that set the tinkerer's heart to joy) into plain vanilla. Have no doubt, I have a long Cisco pedigree rooted first in network engineering, and I recognize the infrastructure efficiencies and scalability set loose by what Cisco is doing. But when it comes down to it, building even the best vanilla producing machine in the world isn't nearly the same type of experience as making 52 flavors scale. That being said, we all better start liking vanilla a little more.
As Beth Pariseau recently pointed out
, Brocade and Cisco's numbers seem to indicate what might be an interesting (but hard to substantiate) story. The story is one of Cisco suddenly taking a bigger bit of the FC switch market, and everyone is asking why, because they think the MDS is a weaker product than Brocade's DCX offering. In turn, they are concluding it is because of the VCE coalition with EMC and VMware that is dragging Cisco into storage deals. For my part, I'll call it like I see it - I think this is just the tip of a trend we're going to see burst through the surface over the next couple of years, but the implications are a little bit bigger, and the other vendors better be worried.The real story starts here:
I think MDS is being carried along by UCS, irrespective of the VCE coalition. Moreover, MDS is demonstrating that UCS is making the specs and competitive differentiation of the other components deployed along with UCS completely irrelevant.
It has long been the case that you could design for the needs of any piece of competent hardware as long as you could design the total surrounding infrastructure. Reality has been that we couldn't control the surrounding infrastructure, couldn't predict the future, and often had to compensate for this. We compensate for it with our own careful design and guesses, and by integration of parts with state of the art specs. Therein was the art of infrastructure design. With UCS, say "bye bye art", and say "hello consumption". Cisco is delivering a package of infrastructure that is end to end designed, and where the future is largely predictable and well known. At that point, the differentiation of the individual components no longer matters.
This is one of the compelling, but perhaps under recognized, value propositions behind both VCE and UCS. When you're turning to converged infrastructures and comprehensive infrastructure stacks like within these offerings, issues like backplane bandwidth, whether real or perceived, are much more easily dealt with to the customer's satisfaction. VCE and even UCS sales are almost always fully contained, completely designed solutions where all these factors have been taken into account. Moreover, with vBlocks, or even sets of UCS, the customer has a clear view of the future of how they can grow an environment and make sure everything stays well aligned.
For the VCE customer, it looks like the first infrastructure has arrived where the infrastructure is well built to start with, and will stay that way over time. It might be vanilla, but it looks like the best vanilla around, and you can consume as much as you want without getting unhealthy. In turn, differentiation around internal speeds and feeds may be put aside - much like an Apple iPhone or iPad, where the experience is delivered by the total integration, and where internals are not even a consideration.
Time will tell whether Cisco can keep delivering on this vision in the face of changing business demands, but if they do, it is reasonable to suspect VCE and UCS will permanently change the vendor landscape around all aspects of server and data center fabric hardware. Consider MDS sales an early indicator of the possibilities.
I'll revisit UCS soon in a little more depth, because recent discussion with fellow analysts, end users, and Todd Pavone at EMC World (who told some compelling stories about just how much you could build out of vanilla - you might say he's a toppings and accoutrement specialist) led me to believe that the core differentiators still aren't really understood (isn't this the case with most Cisco products?).