Pages Navigation Menu

Army won’t budge in fusion software debate

The spat over whether the Army should be forced to embrace Palantir's link analysis software could be moot in three or four years. By then, intel leaders hope to have transitioned away from multibillion-dollar licensing deals to online apps and secure data clouds.

Demonstrating the DCGS-A fusion system for members of Congress

In that case, it wouldn’t be up to a lawmaker or general to choose which software a military analyst should rely on to diagram the links among tribal leaders or suspected insurgents, which is what Palantir does.

The analysts would select from multiple apps vetted for security and accuracy. The software company that makes the best apps – those that get the most use -- would make the most money under a metering scheme still to be worked out.

It’s an interesting vision, but with thousands of U.S. troops still in Afghanistan and an unpredictable young leader in North Korea, the present controversy over Palantir must be dealt with.

Educating Duncan Hunter
The Army launched a PR offensive this week to try to keep Rep. Duncan Hunter (R-Calif.) and his allies in Congress from forcing the Army to incorporate Palantir into its global intel fusion toolkit, known as the Distributed Common Ground System-Army. Hunter's back and forth with Army Chief of Staff Gen. Raymond Odierno at a hearing last month went viral on the Internet.

Palantir’s fans say the company's product is more intuitive than others and that its popularity among analysts at CIA, the Marine Corps and special operations forces proves it. The Army says the problem with Palantir is that it’s built around a propriety Palantir code that makes it hard to share information.

Sure, PowerPoint visualizations can be sent around, but these are “flat files” that can’t be changed or embedded into other products, Col. Tom Miller, the intelligence chief for Army special forces, told reporters.

Lt. Gen. Mary Legere, the Army’s top intelligence officer, was emphatic that the service won’t buckle to pressure from Hunter or Palantir. She said other software suppliers “sacrificed profit and developed new models” to help solve the intelligence community’s information sharing problem, and that Palantir must do the same.

She offered business advice for the Silicon Valley company: Embrace the transition to cloud computing, apps and software standards or be left behind.

“I believe that they’re a really smart company. They’re going to make those adjustments, and we wish they’d make them now,” she said.

<< Full transcript available under "Agencies" in the Vault >>

Two of Palantir’s competitors already have link analysis tools in DCGS-A. They are Analyst’s Notebook by IBM’s i2 unit, and Axis Pro by Austin-based Overwatch, which is part of Textron Systems.

Palantir, which did not answer an email inquiry, has been trying for two years to lobby – some say bully - its way into wider use by the Army. The Army has used Palantir a little bit as a rush-to-the-field quick reaction capability. Getting it into DCGS-A would be a huge win for the company.

So far, Palantir has succeeded only in muscling its way into a cashless cooperative research and development agreement with the Army lab at Aberdeen Proving Ground, Md. Engineers haven’t been able to figure out how to make Palantir and DCGS-A play together.

“It’s not a trivial change or trivial problem that we’re trying to work through. It actually requires some fundamental adjustments to the data structure,” said Col. Charlie Wells, the service’s DCGS-A program manager.

More than link analysis
The truth is that the Army is a bit baffled by all the hubbub over link analysis. It’s critical in places like Afghanistan, but it’s not so important in a place like the Korean peninsula. Analysts use DCGS-A and its numerous software tools to glean intelligence from all kinds of collections, from signals intelligence, to drone video, to human intel reports and satellite imagery.

The feeds arrive at Humvees equipped with antennas and processing shelters. From there, the data’s sent to analysts -- typically non-commissioned officers working in high-tech tents known as tactical operations centers.

As part of its PR offensive, the Army set up a simulated TOC on the lawn just outside the Fort Belvoir, Va., headquarters of the Intelligence and Security Command, which manages DCGS-A.

Reporters were given a tour Thursday complete with a simulated intelligence production scenario. Analysts crunched simulated collections and visualized them on laptop computers and large displays. The goal was to give a commander the intel to decide whether to “remove” a character named Mohammed from the battlefield.

Lawmakers and staffs were on hand the day before, and Hunter himself was scheduled to visit Friday.

More intuitive
Palantir was a hot topic during the media day, but the Army also laid out its plans for the future of DCGS-A. The Army has started equipping deploying forces with a new operating system called Hunte, which is supposed to be more user friendly and intuitive than the current Griffin version. The software is named for an Army specialist who died in Iraq. The Army accelerated development of Hunte in response to criticism from Hunter, who has a special interest in the troops as a Marine veteran of the wars in Iraq and Afghanistan.

The Army also has two pilot projects in the works. One, called DCGS-A LITE will be a portable version of the software and computers for special operations forces. The other pilot will test a new version called DCGS-A Cloud. Analysts are scheduled to start testing the pilot version at regional theater brigades by the end of this year or early next, with fielding scheduled for late 2015 or early 2016. DCGS-Cloud is supposed to mesh seamlessly with the Intelligence Community Information Technology Enterprise, known as ICITE.

Director of National Intelligence James R. Clapper said last month that some in the community are resisting ICITE’s roll-out. It’s entered the “passive aggressive phase,” he told an audience at the C4ISR Journal conference.

The Army's message was that it's not them. Legere said over and over again that the Army is getting onboard with ICITE.

If all goes as planned, agencies will pool their apps and their data, subject to classification rules. In theory, a military analyst could choose to use an app supplied by the CIA or another agency. There'd be no need for lobbying, because the decisions would be in the hands of analysts.


  1. The Army wouldn’t need to put so much time, effort, and money into PR events for DCGS if the program actually worked. If DCGS actually works, then why didn’t they showcase the version that is fielded to the troops in Afghanistan?…it doesn’t actually work so they have to stage these smoke and mirror dog and pony shows to make it look like it works, for fear carers will be ruined if they admit it is a failed program and a waste of money.

    • DCGS-A is a SYSTEM of plugins. Microsoft at one time offered to Build a unilateral system that would encompass Army requirements. The Army obviously didn’t go that route. The Army went with multiple companies and then adopted a SYSTEM (DSGS_A) that provided a common thread to link up all the systems. DSGA-A does work. The systems functionality is based on a competent Unit S-6 with a systems integrator. The real problem is the Army has a way of throwing warm bodies at positions with no prior training. (A cook one day, a Systems Admin the next)

    • SOF Task Forces use DCGS-A and Query Tree just fine. The Palantir GUI takes, no joke, 51 minutes to load via tactical satcom. Thats on a site with the biggest pipe. I know one TF with a 4k bps link and I can’t even imagine how long it would take there.

      You can be up and running with minimal training on DCGS tools. I know people who’ve worked with Palantir for some time and still find it daunting. If you just build link analysis, Palantir is a waste. Analyst Notebook does it just fine.

  2. The Cloud-enabed “pay-for-use” SaaS model is quickly becoming a reality. The Army and other agencies need to embrace this model to bring down high license and maintenance costs. Get on board or get left behind!

  3. The DCGS-A is not a system, it is a compilation of systems and data. It does work. 1. You do have to be trained…. but isnt that the truth for anything you work with? 2. Connectivity is a must for reaching out to data servers. There are so many programs with so many capabilities in DCGS-A that is like comparing apples and oranges with Palantir. Palantir is a single software system with one output. DCGS-A is a capability, not a single software system. These articles are very misleading. How do I know? I am DCGS-A subject matter expert. And I am not biased. Palantir is a good product for what it does.

Leave a Comment

Your email address will not be published. Required fields are marked *