The Army’s intelligence and IT communities have spent the past several years building a family of systems designed to remedy past intelligence failures that were caused by an inability to share information between systems and agencies. That system is now up and running, but the Army is mounting an all-out campaign to defend it.
After having struggled through the earlier years of the wars in Iraq and Afghanistan with intelligence-gathering systems that could not easily interoperate with each other, DoD’s stance is essentially, “never again.” The Army, the largest player at the moment in the battlefield-intelligence business, says common standards are an absolute must for any IT solution going forward. Proprietary or stovepiped systems are forbidden, even if that means the user experience might be less than optimal in the short term.
“Sometimes we have to explain that access to data and adherence to the intelligence community’s standards may be more important than the thing that seems easier, but creates issues,” said Lt. Gen. Mary Legere, the Army’s deputy chief of staff for intelligence.
Legere and other top Army intelligence officials are spending a full week trying to detail that explanation to members of Congress in a series of walk-through demonstrations of the Army’s Distributed Common Ground System (DCGS-A), the end-to-end collection of integrated IT systems it developed beginning in 2007 for intelligence collection, analysis, dissemination and tasking. On Thursday, leaders took a small group of reporters on the same tour in a rare opening of the gates of the service’s Intelligence and Security Command headquarters at Fort Belvoir, Va.
Heated exchange on Capitol Hill
Army leaders freely acknowledge the no-compromise stance toward data standards creates a fundamental tension between the interoperability imperative and the need for systems that might be able to deliver user-friendly battlefield capability sooner.
One such system, a commercial data-analytics product called Palantir, is the poster child for that tension, which boiled over in a House Armed Services Committee hearing three weeks ago.
Rep. Duncan Hunter (R-Calif.), a former Marine who says he’s heard from numerous soldiers who want to use Palantir and have been turned down, spent three minutes criticizing the Army for portions of DCGS-A that he said duplicated capabilities already available in commercial industry, then began to leave the room without waiting for a response from Army Secretary John McHugh or Chief of Staff Gen. Ray Odierno.
“First of all, I object to this,” Odierno replied angrily. “I am tired of somebody telling me I don’t care about our soldiers. I’m tired of these anecdotal incidents.”
“General, you have a very powerful personality, but that doesn’t refute the facts that you have gaps in the capability in the system the Army’s using right now,” said Hunter, retaking his seat.
“We have more capability today in our intelligence than we’ve ever had,” Odierno shouted over him. “A company commander today with DCGS-A has 20 times the capability that I had as a division commander in 2003. I want you to get briefed on DCGS-A.”
Data instantly accessible to intelligence community
Hunter was scheduled to take part in the walk-through tour on Friday, the Army said. Assuming he’s given the same talking points that Army leaders delivered to reporters the day before, he will be told that the system has met every one of its milestones to date and that the Army is aggressively pushing forward with a more user-friendly release of the system’s software. It’s called “Hunte,” after a soldier who was killed in the early days of the Iraq war.
DCGS-A is the largest of a broader family of systems that have their own variants in the Navy, Marine Corps and Air Force, but the Defense Department says the essential feature of all of them is that they enforce common data standards promulgated by the Office of the Director of National Intelligence. That means every bit of data the system collects, ingests or analyzes is potentially available instantaneously to any member of the intelligence community with the proper clearance, and isn’t obstructed by the need to translate it into a format that another system can understand, or that a human being will have to manually type into another keyboard.
“The Army’s never dictated these standards itself, but as the force that provides the global backbone for combatant commanders around the world, we really cannot afford the inefficiency of the one-off solutions,” Legere told reporters. “We started changing that from 2005 to 2007 with great help from industry, who sacrificed profit and developed new models. But the hard migration handwriting is on the wall for anybody who might have that [proprietary] approach.”
The “hard migration” will come about as a result of the entire intelligence community’s pending migration to a common cloud-based architecture. The Intelligence Community Information Technology Enterprise (ICITE), still in its infancy, was mandated by Director of National Intelligence James Clapper as a way to reduce redundant IT costs and promote information sharing.
The Army is already piloting a cloud version of DCGS-A so that it’s ready to migrate its intelligence capabilities into ICITE when the time comes, and so that it can influence the final architecture, which officials say will de-emphasize individual software products and focus on the data itself. The data will be viewed and manipulated by a broad ecosystem of widgets and apps available in the IC cloud.
The majority of the software components in the current incarnation of DCGS-A are essentially commercial, off-the shelf, as is most of the IT hardware. The Army says those components all conform to common standards, and if it’s forced to return to the era of proprietary solutions in any way, it will be dragged kicking and screaming.
“The efficiencies that we’ve gained in DCGS-A by collapsing stovepipes and having common hardware and software from the lowest level of the Army all the way up to the highest echelon has brought tremendous savings to the Army,” said Col. Charlie Wells, the program manager for DCGS-A. “We’ve already realized $300 million in real procurement savings. We’ve got a cost-benefit analysis that proves that we’ve also got $1.2 billion in cost avoidance. Anything that takes us back to separate intelligence stovepipes would not support that strategy that’s been so efficient to this point.”
The Army recognizes though that Palantir has a user interface that’s easy to use, which is why many soldiers have taken a liking to it. The Army has entered into a cooperative research and development agreement with the company to try to determine whether Palantir can translate the proprietary format in which it stores its data into an open format that DCGS-A can use.
‘…Where do you believe your secret sauce is?’
However, the challenge is not a small one, said Dr. Russell Richardson, the senior scientist for the Army’s intelligence command.
“A lot of it gets down to where do you believe your secret sauce is,” he said. “Our users see the ease of use of Palantir, and they see the secret sauce as being its data visualization capabilities. In reality though, their secret sauce is much more holistic. It’s how the system, from the data to the visualization is all very, very tightly integrated. The problem with that is that’s what causes the difficulty in having them take their system and adhere to our standards. That’s breaking up their secret sauce, but this is kind of the cost of entry. What we’ve challenged them to do is maintain their secret sauce but work within our architecture.”
Until the system does, any intelligence that’s collected or analyzed in Palantir is essentially walled-off from the rest of the DCGS-A ecosystem, the Army says, making it useful only within the proprietary application itself, unavailable for tasks such as tasking intelligence assets or controlling sensors, and unable to be shared with or manipulated by the broader intelligence community.
Nonetheless, Legere said, the premium that Hunter and many soldiers place on the usability of the system is completely understandable.
“Rep. Hunter has three tours as a Marine. He has a clear understanding of the tension at that forward edge, and he wants that to be easy,” Legere said. “I understand that need too, and we’re doing a lot to work on ease of use. What I’ve asked him to appreciate is the need to marry up two different concepts: ease of use when it comes to access to data and sensor control, and ease of use when it comes to a front-end graphical user interface. If we can get that together and there are industry partners who can deliver that, we have nirvana.”