At EMC’s fall analyst event, Howard Elias, President and COO of EMC’s Global Services organization, quietly dropped a nugget that Michael Capellas was stepping down as the head of the VCE coalition. Surprisingly, other media organizations haven’t written about this.

Capellas was brought into right the ship at VCE or Acadia as it was know at the time in May of 2010. It was a huge move in my view and further evidence that Joe Tucci and his crew have an uncanny knack at hiring great executives.

Capellas was best known as CEO and Chairman of Compaq, which he sold to HP in 2002 for $25B. He of course also headed Worldcom and led the company out of a disastrous accounting scandal. But Capellas brought much more to the VCE table. First he’s technical. Not only was he CIO of Compaq before becoming CEO but also he’s known in the industry as being product and technology saavy.

As well, Capellas likes attacking hard problems. He took over Compaq at a time when it was getting crushed by Dell in the market. He navigated Compaq through a contentious acquisition as HP’s board was divided on the merger. And he took over WorldCom, which was a complete mess at the time, leading it out of bankruptcy and eventually selling it to Verizon for nearly $7B.

Capellas was the perfect choice for VCE. Initially, VCE (as Acadia) was billed as an accelerant to build, operate and then transfer technologies with a heavier service model. This approach was confusing to channel partners and customers. It created problems as well in that EMC, Cisco and VCE all had different code release schedules. Under Capellas’ direction, VCE became much more streamlined and focused on selling and integrating and left the services opportunity for its partners.

Capellas had a clear vision, which he articulated in several conversations I had with him and in interviews on theCUBE. Last May, Stu Miniman and I talked to Capellas. I got the sense at that time that he felt his work at VCE was largely done. That he’d put the organization on the right track. Capellas took us back to the early days of VCE with comments from this CUBE interview:

The history of VCE goes back a long time. I have two parallels. If you think about the world of networking with fixed networks and data networks and point-to-point, etc…then IP networks came along and we said I’m going to write to one place and I don’t need to know or care about the physical layers beneath it. That spawned ten years of extremely rapid technological advancement. Between both Cisco and EMC we said we know there will be a convergence of where the network starts and compute ends – what’s a blade, what’s a switch, how storage fits in and then how virtualization fits. So you had the pieces: compute, networking (the interconnect) storage, virtualization and we said we can develop an alternate model, very quickly, and not be constrained by big company (mergers), and create the next x86 class mainframe by using the best of capabilities. It was an alternate way to do fast innovation using the best technology components with the understanding that, just like networks, there would be convergence.

Capellas as a Cisco board member had visibility into these plans, discussions and directions. He also has a perspective on the competition, mentioning that he had great respect for IBM and their entrance into the market validates the space. I asked him about IBM and its claims to be embedding application knowledge into the system and he commented:

What defines (this space) is you really do have to look at applications and workloads. And people ask what do you use and the answer is it depends on what you’re trying to do…IBM has knowledge of its stack and certainly knows the applications space…(our approach is) we have taken the best in class components from the fabric, blades and storage to virtualization. This is a new way to deploy infrastructure as you need it where you don’t have to know or care about the physical layer beneath it. 

The implication is IBM’s trying to optimize its stack and create value through integration of its components. One could say the same about Oracle and HP. VCE, rather than take a single stack within one company approach, is virtually integrating. This brings the complexities of having to integrate across three companies. In my experience, sometimes it’s just as hard to integrate within a large company. IBM is a good example. In my view it’s taking more time to integrate the storage piece into Pure seamlessly (see my post on Wikibon as to how IBM’s Andy Monshaw will fix this problem.

I love the concept of converged infrastructure. It’s like the mainframe coming back to life in open systems land. I especially like the business model of EMC and Cisco going after the huge TAM of server and storage markets with an integrated approach. There are a lot of questions about the EMC and Cisco relationship, which are reasonable given the Nicira acquisition by VMware (see John Furrier’s Breaking Analysis on this topic). The fact is V-C and E execs will say all the right things but EMC grabbed Nicira through VMware away from Cisco and its creating friction amongst the M&A teams. As well, over time, the software-led infrastructure trend, or what VMware now calls the Software-defined Data Center, puts the two companies on a long-term collision course.

But for the foreseeable future there’s no reason to think this relationship will end. The idea that VCE is losing money so EMC will pull the plug I think is off base. VCE is growing and competing effectively in the market – giving EMC more, not less account control. Cisco has a very nice channel working through EMC’s sales force and at the same time is playing the field with the likes of NetApp with FlexPod. It’s a classic co-opetiton model and makes total sense to me. Partner up where it makes sense and may the best man win.

Capellas put VCE on the path for success. At 58 years old, with tons of street cred, he still certainly has time to do more in the technology business if he so chooses. I wish him the best.