How and why the military should adopt the cloud

At the Association for Enterprise Information (AFEI) 2010 DoD Enterprise Architecture Conference this week, I gave the keynote presentation on (what else?) cloud computing — specifically, how it should work within Department of Defense. The DoD is the poster child for organizations that should be using the cloud, but it must first overcome several issues — and not just technical ones.

The challenge here is to present yet another new, hype-driven concept and sell it within the context of the existing billions of dollars of IT projects that are under way for the U.S. military. No doubt, everyone in DoD IT has heard of cloud computing, but few understand how it could work within their problem domains and support their missions.

The fact of the matter is that the focus has generally been on the mission first, and IT planning and architecture second. So what you get  — and what the DoD, like many large organizations, has — are thousands of mission systems, with no real idea of how they all work and play together, supporting the larger mission. Moreover, the one-system-one-application approach to IT used by so many organizations such as the DoD means that inefficiencies are designed in from the start.

Cloud computing would bring real advantages to the DoD. For example, the DoD has been asked to tighten its belt, such as by consolidating data centers. The cloud can help there. Also, there’s an understanding that agility is now an essential attribute needed to quickly to align to mission changes; that’s a very important concept, considering the events of the last 10 years. The cloud can help there, too.

In a key way, the DoD is well-positioned to take advantage of cloud computing, unlike many organizations. It has done a much better job of adopting SOA than the rest of the government, so many of its existing systems are already treated as sets of services. This allows the DoD to mix and match services on platforms that are best for those services: public clouds, private clouds, or traditional systems.

But what does this mean to the rank-and-file DoD IT staff and leaders?

Focus on componentization before moving to cloud computing. The idea is not to force-fit IT onto cloud computing platforms; it’s about rebuilding the core IT infrastructure as many components parts: data, services, processes, images perhaps bundled into virtual appliances that could be portable among cloud platforms. The idea is to treat these components as items that can be moved to any platform easily, allowing the DoD to run its systems on the platforms that are most efficient and effective, and to quickly align to the requirements of the missions.

Leverage private cloud first. While I often don’t mention this as general advice, the security issues around DoD systems are so sensitive that they can’t live on public clouds — at least for now. The DoD needs to get good at private clouds and move to better and more effective platforms when it can. If it componentized well, the use of private cloud technology should not be much of a challenge.

Bring in new people. I see the same six or seven big government contractors running massive IT contracts, and they have no core, holistic, meaningful strategy around cloud computing — just a few small deployments. Right now, I don’t see enough cloud-savvy talent within the larger “Beltway bandits.” The DoD should consider spending a bit more money on talent from the smaller, more innovative IT service companies that can supplement the knowledgable “Beltway bandits” and lead the rest.

This article, “How and why the military should adopt the cloud,” originally appeared at InfoWorld.com. Read more of David Linthicum’s Cloud Computing blog and follow the latest developments in cloud computing at InfoWorld.com.