DARPA program that could enable JADC2 at risk of slipping through the bureaucratic cracks

DARPA coded a program that can write its own software to build networks on the fly. But the program is at risk of falling through the cracks of the sustainment process.
Software engineers perform paired programming March 25, 2019, at Hill Air Force Base, Utah. (U.S. Air Force photo by Todd Cromar)

Software designed by the Defense Advanced Research Projects Agency (DARPA) that purports to solve a critical battlefield networking challenge is at risk of slipping through the cracks of the military’s acquisition and sustainment system.

The System-of-systems Technology Integration Tool Chain for Heterogeneous Electronic Systems (STITCHES) program— developed to create on-the-fly networks via self-writing software — has broad support among the military services, but no one can figure out how to pay for it, a DARPA official told FedScoop. The software offers commanders the ability to link data from disparate platforms like different weapons systems — a capability senior leaders many have been clamoring for to connect “every sensor, every shooter” under the Joint All Domain Command and Control (JADC2) concept.

The problem is no one can figure out what “color of money” to use for the program. It’s caught between the acquisition authority for research and development and that of sustainment. (There has been a recent push to give software its own color of money, or “budget activity” in legal terms, but has yet to materialize in regulations.) On top of this, DARPA’s funding for the program will run out in the spring after it already used an extension on the program.

“We are getting tremendous pull and demand for it,” Tim Grayson, director of the Strategic Technology Office at DARPA recently told FedScoop. “The thing that has been the challenge with STITCHES is it’s weird.”


When the money runs out in spring, the STITCHES code will remain available to “appropriately authorized personnel” in the DOD open source community. But the expertise of those that built it will dissolve when the program expires.

“Anyone can develop their own STITCHES graph-based database but the current knowledge of the team which generated the existing database will be disbanded,” Grayson said.

A JADC2 enabler

STITCHES promises to deliver the type of data-sharing capabilities at the heart of the new concept of JADC2. The idea is to build a military Internet of Things by linking data from operations and hardware in air, land, sea, space and cyberspace, and funnel that data to commanders and artificial intelligence-enabled machines for better decision making. STITCHES could provide that data link without cumbersome data standards for endpoints by spinning up its own data links and interoperable networks on the fly.

The system was tested in one of the large-scale “on-ramp” events hosted by the Air Force in September. DARPA said STITCHES was able to link different platforms that were built decades apart, allowing for interoperability and data sharing. The software is also entirely DOD-owned, without any commercial proprietary tech that could cause further snags in implementing it.


“The toolchain does not force a common interface standard; rather it rapidly creates the needed connections based on existing fielded capabilities obviating the need to upgrade in order to interoperate,” according to DARPA.

There is also no unit structure set up to effectively use the software, Grayson said. “STITCHES falls in the seams between the boundaries of existing systems and System Program Offices.” The “middleware” that sits between backend systems and ends points needs “geek squads” or “Software Engineering Squadrons” to fully implement, he added.

“The Department is moving in this direction, but currently the STITCHES end user does not fully exist,” Grayson said.

Latest Podcasts