Looking for a silver lining in military clouds

Two significant events with a bearing on military cloud computing occurred on the same day in the latter half of April.

In the first incident, Amazon’s cloud computing network went off-line for more than a day, taking with it hundreds of businesses that had taken an early leap into cloud services for everyday applications. No cause has yet been identified, though at least Amazon has the resources to eventually fix whatever caused its cloud failure.

Later that same day, Defense Systems covered a speech in which Defense Department CIO Teri Takai cautioned against jumping too quickly into cloud computing. “If we move to a cloud environment with today’s technology, it would make the world worse, not better, from an enterprise perspective [and] from a security perspective," she said. "We’re still at Stage 1.”

So on the one hand, there is Federal CIO Vivek Kundra quickly moving forward with a cloud-first policy, even requiring every federal agency to identify three applications or systems that will move to the cloud in the next year and a half. On the other hand, some people, like Takai, who are responsible for their department’s IT strategy have now popped the parachute to slow down the momentum.

Nobody expects brigade combat teams to access their battle applications via the cloud any time soon. But a significant portion of the long-term cost savings and efficiencies that DOD was expecting to materialize via cloud computing and data consolidations will be longer in coming.

IT managers are furiously dissecting what they know about the Amazon cloud failure, and the clear lesson learned so far is the need for a backup system that kicks in immediately and transfers applications and data to a secondary system. It needs to be transparent to users so they don’t know if they are accessing the cloud or a backup system, much like a soldier doesn’t care if he’s communicating via wireless or satellite as long as the connection goes through.

There’s a mantra in aerospace and defense design that no single point of failure should take down a vital system on which people’s lives depend. In other words, an airplane shouldn’t crash because one system fails. It should be the same with military cloud computing. There will always be failures when technology is involved. But in the case of Amazon’s cloud, it was not the technology that let down its users; it was a failure of the company to have a backup in place when the inevitable happened.

As Takai said, DOD is only at Stage 1 of its cloud strategy. But now that the heady rush toward cloud computing has been tempered by the first major failure of a cloud, the opportunity exists for the military to more clearly define a strategy going forward.