Top Pentagon officials are defending the decision to award a potentially multibillion-dollar cloud contract to a single vendor, pointing to the current state of the commercial marketplace, existing acquisition laws and battlefield requirements as critical components.
“What [troops] have today to work with … is very limited and that puts them in a sub-optimal situation to execute the next move of their mission,” John Gibson, the Department of Defense’s chief management officer, told reporters at the Pentagon April 23. “We know there is a cost to operating multiple clouds. Not only is there a degradation of services provided — the accessibility of data, the speed of access — there’s also a financial cost to … making the data work in the best possible.”
“The elephant in the room” is the choice to go with a single-award, indefinite delivery/indefinite quantity contract for the likely massive Joint Enterprise Defense Infrastructure cloud, billed as the future go-to vehicle for military data needs for as long as the next decade.
“There’s no question that machine learning and the ability to process data at machine speeds is going to be a critical component of the way war fighters of tomorrow fight. Arguably it is a critical component of how the war fighters of today should be fighting,” Robert Daigle, director of cost analysis and program evaluation at the Pentagon, said. “In order to get there, by definition you’re not buying a strictly commercial off-the-shelf product.”
That global access is a chief reason for looking toward a single provider, Daigle said. Security is another, and DoD officials don’t want the burden of managing the seams between multiple clouds and multiple data “ponds” that would come with multiple providers. Instead, they’re looking to get the military’s widely scattered and heterogeneous data homogenized into a system that can more easily be accessed far and wide by hundreds of thousands of users.
Then there is the current acquisition system. Giving the example of migrating approximately 133,000 DoD unclassified operating systems, each requiring a roughly three-month task order, it would take more than 400,000 man-months to do that under a multiple-award IDIQ contract, Daigle said.
“Think about that from the point of view of a war fighter, when we have somebody sitting at a [forward operating base] … in today’s environment it takes months and months to get access to a server and storage,” Daigle said.
“We’ve not heard anybody say that a multiple-cloud solution is a better solution for providing that capability to the war fighter, and that’s why we’re saying that based on where technology is today, based on where the offerings of the commercial cloud providers are and based on current acquisition law, the department’s optimal solution is a single award contract.”
Gibson and Daigle downplayed the idea that JEDI is the be-all, end-all cloud contract for the entirety of DoD.
“Folks are trying to present this as the Department of Defense moving to a single award for all cloud computing,” Daigle said. “This contract, even if we max out the annual spend, represents about 16 percent of our hosting costs currently, which represent about 12 or 13 percent of our overall business system IT costs. So really the amount that we’re talking about here is not the end of history; it’s one cloud contract.”
Daigle declined to put a potential dollar figure on the program, and Gibson emphasized that JEDI is just the first step in a long journey toward achieving the vision of universally available access to data on-demand.
“If you look at where we are, there is not some silver bullet where you will design something, flip a switch and then it all just happens. So we have to start this process,” Gibson said. “We don’t know the ultimate end state. We do know a single environment is much better than the federated environment we have. It’s about getting started, getting going and we are very comfortable that this will put us in a much better place.”