For decades, the U.S. military relied on its dominance in the information domain to maintain its advantage in all the others—air, land, sea, space, and cyberspace.
Better intelligence meant better situational awareness. By collecting more accurate, timely data, and more rapidly delivering it to those who need it most, U.S. forces could both out-think and out-fight its enemies.
“If you can collect, share, and flow more data to more users, commanders can then make better decisions. It’s a substantial advantage,” says Bill Conley, a former director of electronic warfare at the Pentagon and now the chief technology officer at Mercury Systems.
Increasingly, however, that historic advantage is coming under pressure. Adversaries have watched U.S. operations unfold in recent years and are catching up.
Last February, Air Combat Command boss Gen. James M. “Mobile” Holmes described just that at the Air Force Association’s Air Warfare Symposium. “We know that China has watched what we do for the last 15-20 years and has tried to come up with ways to counter that,” he said. Today, he added, sophisticated adversaries “can threaten our ability to operate in the domains we have traditionally dominated, not by challenging us in those domains, but by challenging us in this information domain.”
Winning in this new world, maintaining the U.S. decision advantage, doesn’t just mean gathering and disseminating data faster and more widely. It also means being quicker to turn that data into information — and then using that information to make better decisions with greater speed and confidence by leveraging the power of artificial intelligence.
“You win by operating at a tempo [the enemy] can’t keep up with and by putting them on the horns of multiple dilemmas,” said Holmes. “We want to create enough options for our commanders that the enemy doesn’t know where we’re going to come from next.”
Conventional operations control each warfare domain separately. Air operations centers control air operations, while ground commanders control land forces. “[But] it takes too long to have air command and control and ground command and control and Navy command and control to get together and decide what to do,” Holmes said. “You have to put it all together.”
This combined approach is now known as Joint All-Domain Command and Control, or JADC2. By eliminating conventional stovepipes in favor of a more integrated and truly networked force, a joint command and control system can connect any shooter to any sensor at any time.
Preston Dunlap, chief data architect for the Air Force, calls it “The Internet of Military Things.”
Enabling this combat-zone military internet are critical advances in information technology, from onboard sensors and data processors at the forward edge of the network to the power of commercial cloud computing and artificial intelligence. Mercury Systems produces the embedded systems and technologies that can leverage those capabilities and bring JADC2 from laboratory to combat reality.
“If you look at the equipment the U.S. military is using, our technology — our RF and microwave components, our signal processing, our embedded computing — is underneath the hood of nearly every radar, nearly every C4ISR system, nearly every EW system,” Conley said.
Like the commercial world, these systems are rapidly evolving. Software-defined technologies are revolutionizing how capabilities are delivered. Instead of building each system in hardware, more powerful technology, based on open systems architectures, can now implement evolving capabilities using software, rather than specialized hardware. Not only does that help to seamlessly shift from one network or technology to the next, it also allows for rapid, iterative change and improvement. Leaping across huge swathes of spectrum like that would have been unimaginable even a decade ago.
Changing frequencies on a radar or radio no longer means having to replace all the hardware, Conley said. “Now we have the ability to electronically reconfigure a piece of equipment via the network,” he explained. You can use it for a very different purpose than may have been the original purpose of that system.”
Rather than thinking of a system as a radar, Conley said, “Now it’s a sensor: It’s a collection of apertures. It’s a collection of signal processors. You can use that same equipment to gather and process multiple data streams in very different modalities to accomplish different tasks. And you can share that data, too, with other platforms.”
Whole New World
As the technology world advances, so, too, does the pace of change. Technology is developing faster than the U.S. military can buy and deploy it in weapons systems. Driven by commercial product cycles counted in months, technology can’t stand still waiting for military acquisition systems that track time in years and even decades.
The disparity threatens the long-term success of JADC2 by creating potential nightmare scenarios in which the military locks in technology choices only to see them block future innovation months or years down the road.
The pace of advancing technology has long been defined by Moore’s Law, which says the number of transistors on a chip will double every two years. That, in itself, was a challenge to military systems that are designed over decades and fielded for decades beyond that. In a software-defined world where machine learning and artificial intelligence is driven by algorithms enabled by silicon chips, capability is accelerating.
Today, AI capability is doubling every three-and-a-half months, according to data from Open.ai. In other words, a system designed on Jan. 1, 2020, will be only about one-tenth as capable as one designed a year later. After two years, it is only one percent as capable. And after a decade — the time it typically takes for a military platform to move from drawing board to fielding — that system would have only one billionth the capability of a newly designed system.
“That’s the challenge we’re up against today,” Conley says. “For JADC2 to be successful, we have to avoid that trap. We can’t lock in or we’ll fall behind.”
To avoid that fate, the systems that enable JADC2 must be built with open, modular architectures and based on open standards — enabling near-continuous upgrades from competing vendors and agile software development methodologies that ensure rapid product improvement.
Dunlap, Air Force’s chief architect, compared those modules — datasets, applications, and other software products — to Lego blocks in a recent presentation for the Air Force Association’s Mitchell Institute for Aerospace Studies. “They need to snap into place with each other to make this all real,” he said. So the DOD must “take its cues from the way the commercial sector has moved quickly and adapted to the digital world over the last 20 years.”
Holmes has famously compared JADC2 to ride-sharing services like Lyft and Uber. “They match riders with drivers,” he said. “We talk about matching sensors with shooters with targets.”
The seamless, real-time integration that ride-sharing software offers with other apps like Google Maps shows the modular future for U.S. military systems, Dunlap said. “Underpinning all this is digital engineering: Open architecture and open standards that ensure those Lego blocks — just like real Legos — actually snap together and work.”
Mercury’s Fellow Systems Architect Matthew Alexander says the Air Force has been following just such an “open systems approach” for years, driven by the recognition that open standards developed in collaboration with industry pave the way for more choice and innovation down the road.
“It’s a multi-tier approach,” Alexander says. At the lower tiers, the sensors themselves are the focus. The Sensor Open System Architecture (SOSA) supports new capabilities for deployed equipment “much faster than was possible in the past,” Alexander said.
The middle tiers enable payloads and services on the same platform to share data seamlessly. The Open Mission Systems (OMS) standard adopted in 2014 allows sensors and weapons systems on an avionics bus (like the ubiquitous MIL-STD-1553) to communicate machine to machine.
Finally, the highest tiers establish interoperability and seamless machine-to-machine communications between platforms and across domains.
Alexander said these defense standards must leverage commercial standards and diverse working groups that bring together industry partners and competitors to achieve the best possible technical solutions with the widest possible adoption. Crucially, these systems must also be “backwards compatible” to ensure legacy equipment is not locked out of the future JADC2 capability.
“The DoD has wisely adopted an approach where you can have different levels of compliance with these standards,” Alexander explained. “Rather than saying, ‘You shall be compliant with every single requirement in this 2,000-page document,’ which would raise the barrier to entry too high and might make it impossible for legacy systems to comply, they lay down minimal compliance requirements as an initial baseline and then add more compliance requirements at higher levels.”
The result, Conley says, is a flexible framework for both vendors and their military customers, enabling the services to manage complexity and costs at the same time. “You don’t want one standard to do everything because it results in gold plating every solution, and that can become prohibitively expensive,” Conley says. Rather, the idea is to tailor the requirement to the need.
Naturally, thresholds for compliance and requirements vary from one use to another. The demands for a supersonic fighter aircraft are higher than for a car operating at less than a tenth the speed, or for a mobile phone: A dropped call on a mobile phone is inconvenient. A dropped connection in battle might cost lives.
“Obviously, the networking piece of this is challenging,” said Conley. “You can’t afford bottlenecks.”
“Call it the forward-operating cloud, or the tactical cloud or even the edge cloud,” Conley said. The operating network in battle has to operate seamlessly, at speed, without putting the user at risk. “There will be data that we get all the way up into the global cloud,” Conley said, “and there will be data that we get into the forward-operating cloud.”
An F-35A boasts some of the most advanced sensors available, as well as onboard processing to analyze and present that sensor data to its pilot. But only some of that data will be sent from the aircraft to other platforms in real time. Once back on base, the maintenance crew can download other operational data and share that as bandwidth and other demands allow.
“As we move into the future, the architectural barriers begin to break down and it’s going to become — pun intended — very foggy,” Conley says. “It won’t be easy to discern where the cloud stops and where the edge begins.”
This is where AI comes in. Pilots, as Conley puts it, “have a day job.” Already fully occupied flying the plane and engaging targets, they would benefit from AI working in the background, analyzing the huge data haul from the aircraft’s advanced sensors, and prioritizing what information the pilot could need and what can or should be shared with other nodes on the network.
In WWII, Britain’s Royal Air Force, though outnumbered five to one by the often superior German Luftwaffe, maintained air dominance thanks to the Dowding System, a real-time reporting structure that funneled information about the location of German planes from radar stations and lookout posts all over the country to Fighter Command Headquarters. There it was rapidly disseminated to local commanders who deployed intercepts to challenge the German attackers.
Back then, sharing information by telephone and recording it with grease pencils, allied air forces ushered in the information age. Today, the digital revolution continues to accelerate the decision cycles.
Yet, as Holmes notes, “it takes more than new technology to make a fundamental change in the way we fight: … A real revolution in military affairs requires a change in the way we think, as well.”
JADC2 aims to do just that. The starting point must be in how systems are engineered, not just how they’re used in battle, Mercury Systems’ Conley says. Augmenting the digital data flow with AI and other tools demands a new, adaptive approach to system design and engineering that can allow for and support the rapid release of system changes and updates as new capabilities emerge.
The spoils of victory will always go to the side that can dictate the terms of an engagement. The more one knows, the greater the advantage. Maintaining the U.S. decision advantage and the deterrent power of superior military forces is not an option, but a necessity, Conley notes. Ensuring those advantages must start with the decision to adopt open standards today.