If the Pentagon Drops JEDI, Then What?

The outlook for the Joint Enterprise Defense Infrastructure contract—the Pentagon’s $10 billion general purpose cloud project—looks gloomier after a federal judge denied a motion to dismiss political interference allegations made by Amazon Web Services. 

Since January, when the Defense Department sent Congress and media an information paper suggesting such a decision would likely imperil the contract, officials have been repeating the same line: The capabilities meant to be provided by JEDI, which was awarded to Microsoft for a second time in September, are still urgently needed.

“We continue to have, I think you know, an urgent, unmet requirement for enterprisewide commercial cloud services at all three classification levels, but we remain fully committed to meeting these requirements, we hope through JEDI, but these requirements transcend any one procurement,” DOD Press Secretary John Kirby said during a May 10 briefing. “And they’re going to have to be met one way or the other.”

About a week before Kirby’s statement, Deputy Defense Secretary Kathleen Hicks declined during an Aspen Security Forum event to comment on the litigation but tied moving to a cloud architecture to the future of innovation at the department. Hicks said officials will have to assess where it stands given the ongoing litigation and then “determine what the best path forward is for the department.” 

But with cancellation largely assumed to be coming, what the best path forward might be isn’t clear. In the years since the JEDI concept emerged, cloud technologies and the DOD cloud landscape have evolved. 

Alex Rossino, a senior principal research analyst for Deltek who tracks federal IT market trends with a focus on DOD, told Nextgov one of the biggest changes since the JEDI saga first began is an increase in the number of compliant cloud solutions. 

Perceptions around the security of cloud have changed, too, such that now the cloud is viewed as more secure than what agencies can provide because vendors are on top of the latest technologies, Rossino said. 

“Agencies have shifted the burden for resource security onto vendors, as opposed to maintaining themselves,” Rossino said. “So that's a big change, and when that changed that opened the door to all kinds of all kinds of procurement and you can see a huge rise in cloud numbers starting in 2017.” 

According to Rossino’s analysis of spending data, DOD awarded nearly $400 million more in cloud contracts during fiscal year 2020 than it did in fiscal year 2019—and that’s not counting the JEDI award. 

DOD officials themselves have emphasized the JEDI standstill hasn’t brought other cloud efforts to a halt. In a recent interview with Nextgov, DOD’s Deputy Chief Information Officer for the Information Enterprise Danielle Metz said the military services and the Defense Information Systems Agency are helping plug the general purpose cloud hole. 

“I think that we're achieving an aspect of what we meant by the enterprise general purpose cloud concept,” Metz said.

Still, Metz and experts including Chris Cornillie, a federal market analyst at Bloomberg Government, who told Nextgov that even as more clouds pop up at DOD, the agency still needs an enterprisewide capability to stitch everything together. 

“While all of the different military services are building their own platforms, what we’re risking is just different stovepipes,” Cornillie, who recently wrote a report outlining what might happen if DOD cancels JEDI, said. “There still is no single enterprise cloud that can share information across all aspects of DOD.” 

Cornillie pointed to programs like cARMY and Cloud One, the cloud computing environments for the Army and the Air Force, respectively, as well as the Joint Artificial Intelligence Center’s Joint Common Foundation, as examples of “stopgap measures.” There’s also milCloud 2.0, DISA’s on-premise cloud, which is the required home of some defense agencies and DOD field activities workloads.

Many of these capabilities are still developing, but Cornillie said they’ve been successful so far because they’ve embraced commercial best practices and also given DOD customers access to both AWS and Microsoft cloud services. Cloud One and cARMY use both AWS and Microsoft’s Azure, and General Dynamics Information Technology, which now supports milCloud 2.0 after acquiring the original awardee, signed an agreement that will allow users to access general-purpose cloud services via AWS through the milCloud 2.0 contract. 

“But I think that the successor will be bigger,” than any one of these options, Cornillie said, citing the Central Intelligence Agency’s C2E contract, a project often held up as a model for major cloud contracting initiatives. 

But others suggest that cloud at DOD—and cloud technology in general—has changed such that alternative solutions may be more suitable, or at least that the business case for a JEDI redux hasn’t been clearly articulated. 

Stan Soloway, chief executive of Celero Strategies and a former DOD acquisition official, told Nextgov the need for an enterprise cloud capability hasn’t changed for DOD, or even for other large enterprises. But how entities are “getting to the answer” has shifted. 

“The question of whether you want or need one single contract for this massive enterprise, bigger than any other enterprise in the world, I just don't see it anymore,” Soloway said. “There's just too much. There's too much dynamism in the marketplace.”

Soloway said he never thought a contract the size of JEDI made sense in the first place because some components were always going to go their own way. There may well be a justification for replacing JEDI with a similar framework, Soloway added, but in light of technological changes and the growth of other options within DOD, he’s just not sure what that justification is. 

For his part, Rossino posits there may be a faster, more secure way to provide the cloud capabilities needed to support Pentagon priorities like artificial intelligence, multi-domain operations, and Joint All-Domain Command and Control, or JADC2. Rossino said a data hub, which would pull information from various systems into a common environment, makes more sense than another single-award solution. 

“It allows formatted data to be leveraged by advanced analytics and machine learning tools, and that's exactly what DOD wanted to do or wants to do with JEDI,” Rossino said. 

Ultimately, DOD may need to simply improve its ability to explain its requirements. 

“DOD has been terrible at communicating what it wants and why it wants it,” Rossino said. “Because they can't communicate, and continue to not be able to communicate what they want and why they want it, they continue to have these problems.”