Managing the Lego Data Center
During customer briefings, I’ll often use the concept of the Lego data center when talking about the vision behind Data Center 3.0. The joy of Lego (for me at least) is the ability to build something then, (sometimes violently) break it apart and build something else. This is the operational model we look to bring to the data center—sans most of the violence–with Data Center 3.0.
Virtualization has a central role in this vision, but asset virtualization introduces its own challenges. Arthur Cole has a posting today where he notes that “Management issues begin to get more complicated once virtual servers start to rely on virtual storage.” I’d actually take it a step further than that. Even today, I would say that data center virtualization efforts are simply tactical responses to a given problem, say server sprawl. Within a given technology silo, they may fix a problem, but, as a whole, is the data center any better off? Probably not, and managing disparate virtualized assets on top of physical environments might make things a bit worse. I don’t think “virtualization” is going to deliver any net benefits to the data center until we can address the issues of orchestrating services across technology silos. So, if I move a virtual machine, the SAN, firewall, switch and load balancer, etc, all have a clue on what to do next. This should be the next great battle in the data center. There are a lot of interesting solutions out there and, of course, we have some interesting things in the pipeline too. The problem is that, to date, no one has a good holistic solution, so it seems like some M&A activity needs to happen (no, that’s not a hint) or at least some solid alliances built, before customers get something truly practical.