This is the second of a multi-part series exploring why our IT organization is so aggressively transforming EMC’s own datacenters into Private Clouds. In Part 1, I describe how our IT group’s strategy shift began. Despite what you might think, it wasn’t driven by technology.
In this part, we look at what happened to cause our senior IT leaders to second-guess their strategy.
A couple years ago we had a major ice storm in New England. Power outages affected our Hopkinton headquarters datacenter—and our Westborough datacenter. That event forced us to realize we needed to move one datacenter much farther away. Even the best disaster-recovery technologies, products and practices can’t help you when all your sites are taken out by the same event. Needless to say, we urgently needed to address this risk to the business.
Shortly thereafter, the global economy went into a “great crater,” or whatever pundits and historians will eventually decide to call it. In any event, it certainly helped focused our minds when it came to choosing a new site for one of our datacenters. And that potent combination, gaining significant cost savings and mitigating a significant business risk, motivated us to get a datacenter migration project moving and into high gear as soon as possible.
Migrating an entire datacenter is a huge undertaking. In our case, it involves thousands of systems, applications, repositories and other data stores with ever-changing interdependencies and workloads. Orchestrating a move of such magnitude without disrupting the business is no small feat—and fraught with risk.
But it also provides an opportunity. Instead of merely transplanting the Westborough datacenter as is, we can accelerate our datacenter transformation and build a completely new datacenter based on a fully virtualized infrastructure. Most companies in this situation do this one of two ways: move first and then virtualize, or virtualize first and then move. The first option expedites the move, but it delays virtualization’s cost savings and other gains. The second option accelerates virtualization’s gains, but it delays the datacenter move’s benefits.
Meanwhile, the cloud-computing buzz became loud enough to reach our IT organization’s ears. They began to examine cloud as it was then defined, and became worried. Most cloud approaches required a radical re-tooling or replacing of applications, with ripple effects throughout our current and planned system architectures. They’d also mean radical changes in operations, and significant loss of control over service level management and security.
Our senior IT leaders had felt pretty secure in their virtual datacenter strategy. One of EMC IT’s major goals, however, is “architect for the future.” Because it was being proclaimed by many as the Next Big Thing, our IT folks needed to understand whether this cloud thing was really an opportunity they couldn’t yet grasp, or merely another IT marketing-hype distraction.