Cloud Providers: Should You Put Your Eggs in One Basket?
The Defense Department is drawing a lot of buzz in the IT space at the moment for seemingly committing much of its massive infrastructure to a single cloud provider, arguing it’d be too complex to share the burden. But does that strategy make sense for smaller shops, like associations?
The Department of Defense (DoD) doesn’t run its cloud like anyone else, and it just highlighted that with a new RFP.
In the RFP, the agency is asking for a single vendor to fill a contract worth up to $10 billion, for its Star Wars-named Joint Enterprise Defense Infrastructure (JEDI), which aims to put defense apparatuses on the commercial cloud.
That’s, obviously, a big chunk of change for a pretty fundamental part of how the Pentagon works. The contract would represent 16 percent of the agency’s cloud apparatus, according to a Washington Post report from this week. That could be a big win for an industry giant—Amazon Web Services (which won a similar, though smaller, contract for the CIA in 2013) is widely considered the front-runner, though companies like Microsoft, IBM, Google, Oracle, and General Dynamics are also on the list.
Word of this RFP has been floating around for months, and it’s been very controversial the whole time—vendors have been making the case that the cloud should be managed by a number of providers, rather than a single one.
The Pentagon, on the other hand, thinks that the ask is too big to be doled out in pieces and seems to be leaning toward minimizing its complexity, which seems like a reasonable concern for a $10 billion project.
“With the diversity of DoD’s mission, DoD will always have a multiple cloud environment, but we need to do better in applying an enterprise approach to that environment,” Defense Department CIO Dana Deasy wrote in a letter announcing the RFP [PDF]. “To successfully accomplish this, we are looking for an industry partner who will learn with us and help us find the best ways to bring foundational commercial capabilities to our warfighters.”
An Enterprise Approach … Or Is It?
The Pentagon’s stance that minimizing vendors is an enterprise approach is interesting, because the evidence of this stance actually seems to lean the other way.
Per a report from the company RightScale, 81 percent of enterprises rely on a multi-cloud strategy, rather than one service. The company notes it was a decline from the prior year, but considering it’s at more than 80 percent, it’s still far and away a majority.
Another study, from the cloud provider Virtustream, puts the number at 86 percent. Either way, it’s a lot.
(Much of the growth in cloud computing in 2018 is happening on public clouds, by the way; while private cloud services are still out there, RightScale notes, the approach seems to be leaning more heavily on public clouds, or even hybrid solutions.)
Often, the reason cited by many enterprises for this diversity of providers is an avoidance of lock-in. This is a reasonable explanation and one that avoids putting all of one’s eggs in one basket, but what’s interesting is that if you actually ask a vendor, their explanation is closer to the Pentagon’s.
Adrian Cockcroft, the vice president of cloud architecture strategy for Amazon Web Services, told The Wall Street Journal last year [registration] that many of its customers relied on multiple cloud services, but that wasn’t exactly what he recommended:
If you go for a 50/50 split and use lowest common denominator, you’re really hobbling yourself. You’ll go more slowly, because you’re having to train your people twice. You’re cutting your ability to get a volume discount in half, so you don’t get as good pricing because you’ve got two vendors there. What we’re seeing is that predominantly you want to pick one vendor and maximize that footprint.
(This may in fact be the argument that might win Amazon the Pentagon’s business.)
Of course, the problem is a lot more complicated than a sound bite from a leader with a vested interest in growing his company’s offerings—as Forbes notes, the increasing specialization of the various cloud services is a big factor behind the move toward multi-cloud. Ultimately, it comes down to your organization’s needs, and your organization may face issues that prevent a simple split. Maybe you’re locked into another vendor elsewhere—say, if you’re on AWS for your internal data, but a key vendor, like an AMS or LMS provider, uses Microsoft Azure, making it better to have resources devoted to both.
You can take this further, of course: Maybe you use Office 365 for sharing documents internally, making Azure a preferred bet, but prefer Google Apps for email and a few other things. Maybe there are regulatory concerns at play that limit your ability to settle on one provider. Maybe you have a lot of custom legacy code and want to minimize the headache of switching platforms, so you stick with the company that helped you build that integration. Maybe you want your data separate from your software needs entirely, and simply want a cloud provider that can talk your language.
Or maybe there are elements you like about each and don’t want to have to decide.
DoD appears to be going against a wide trend in cloud computing that allows for specialization, in favor of the monolith provider. In many ways it’s forward thinking, but in others, it’s pretty old school. The Pentagon is big enough that it can call the shots, however, and it will be calling many shots with a $10 billion contract hanging over one company.
Your association may not have that level of say—maybe the best you can do is find an approach that works for you, and a vendor that makes sense for your needs.
Or two. Or maybe even four or five. Go crazy.
(Denis Isakov/iStock/Getty Images Plus)