A local coronavirus response functioned as a crucial test of a new data network concept intended to deter Russia and China.
The April 14 briefing contained a warning for the commander of U.S. Northern Command: Stamford, Connecticut, about 56 miles northeast of New York City, was about to become a COVID-19 hotspot. It was a remarkably specific heads-up, and it arrived when only about 1 percent of the U.S. population had been tested, and well before tech companies began working through the complexities of launching contact-tracing apps.
It was a first-of-its-kind warning generated by an experimental data collection system tied into the Pentagon’s planned next-gen combat information effort. Although the data collection and prediction process was unproven, Gen. Terrance O’Shaughnessy and his commanders decided to trust it. O’Shaughnessy dispatched an additional 53 medical personnel north to Stamford — and just in time: Manhattanites fleeing to the Connecticut suburbs helped cause a spike in Stamford’s coronavirus cases: 412 new cases on April 19; 2,109 on April 22.
The episode was an early and critical test of the Joint All-Domain Command and Control, or JADC2 concept. It envisions a system for keeping commanders abreast of complex and dynamic battlefields that is central to Pentagon visions of future warfare. The Stamford success was the fruit of a sub-system developed by NORTHCOM and contractors Apple, Google, Palantir, and geospatial intelligence company ESRI, to help pandemic responders rapidly report information up the chain of command.
“Our team has a better insight for predictive analysis as more of a comprehensive operating picture has been developed,” O’Shaughnessy told reporters on April 22. “JADC2 also helps us know where we should be looking for the next hotspots which is critical for the national approach to fighting COVID-19."
You could be forgiven for assuming that the U.S. military is already pulling data from a variety of sources, pushing it up the chain of command, and then disseminating orders. And it is. But Hollywood depictions of big dashboards in the Pentagon where America’s top commanders micromanage military operations all over the world aren't based in reality. Right now, getting info from the places where soldiers, sailors, and airmen are operating up to the highest level of authority is a slow, indirect process.
“We just didn't have the tools available because we were essentially using tools from several decades ago, and using processes from several decades ago to get information from the edge,” said Lt. Col. Matt Strohmeyer, an action officer with the Commander’s Initiatives Group at NORTHCOM.
“We'll have, let's say, a deployed unit somewhere that's helping out with a medical response, and they will every day get on the phone and they will communicate what's going on, their numbers of ICU beds, or ventilators, or personnel that are sick, and then we'll port that over phone to the next command echelon up,” said Strohmeyer. “That next command echelon up will put it into an Excel file, and then they'll email it up to the next command echelon up, and then they'll compile it in a new Excel file, and that process of what is essentially 1980s technology of aggregating data, that will happen probably somewhere between five to 10 times over at different command echelons, where it finally arrives at NORTHCOM headquarters about 24 hours later. And there the data is then aggregated into a static PowerPoint slide that's presented to the commander, and that means that that data coming from the edge is very old and stale, and it's certainly not real-time at all, and it also is very manpower-intensive.”
That process wasn’t an enormous hurdle when the U.S. military was trying to manage counter-terrorism operations in the Middle East. But they were a real obstacle to rapid response to a pandemic sweeping the nation. NORTHCOM had to push the status quo.
First, they outfitted personnel with devices to allow them to immediately report data that they were seeing into a central location, a cloud. While enterprise-level cloud computing is common in Fortune 500 companies, U.S. commanders have been slower to embrace it, in part because the military isn’t a single monolithic organism. Any particular unit or office might be working with a specific dataset and a specific cloud vendor that no one else is using.
As NORTHCOM confronted the pandemic, there were a lot of players with their own data sets and, in some cases, their own clouds. That data included reports from the medics in the field; geospatial economic, municipal, demographic and other data from ESRI; FEMA and other government data from Palantir; and more. NORTHCOM went to the Pentagon’s Joint Artificial Intelligence Center to house the data from these various partners.
One key aspect: ESRI’s effort to create a geographic picture of what was happening where. The privately-owned technology company based in Redlands, California, is having what you might call a moment in the midst of the COVID-19 crisis. If you go to the Johns Hopkins COVID-19 map, ESRI is a key component. The company provides geographic information services for cities, states and other government agencies for planning. It has a lot of census data to reveal things like income, spending levels, job types, family types, of different places. The data is either open source or comes from their wide user community (with permission). After disasters, different agencies can request specific software or data support to better show how the disaster has affected them, via the company’s disaster-response program.
“The nature of the pandemic is inherently geographic,” one company official said. (The official declined to be named because they weren’t authorized to speak to the media.) The pandemic, he said, generated more requests for support than “all other disasters put together.”
Individual counties and jurisdictions wanted to be able to do their own data collection and reporting on the crisis. The company was suddenly faced with the prospect of helping a wide variety of users handle a massive influx of rapidly changing data from a lot of different sources and in a constellation of different forms.
Federal agencies and even local governments had ESRI tools to organize and map data, but “what was often lacking was an infrastructure for health reporting… Because there isn’t a federal, national reporting system for this sort of thing, which is kind of mind-boggling. Johns Hopkins became a central authoritative source for data related to COVID-19,” said the company official.
The U.S. military combined that with other publicly available data as well as Defense Department information about where military personnel were, what supplies were available, etc. It created an interface that strikes much more closely to that Hollywood vision of digital command and control and much less like a PowerPoint slide, but one that was available to everyone at once, not just the generals in charge. That's when the predictive insights began to pop out.
“We put it all together,” said NORTHCOM’s Strohmeyer, “and then with the help of MIT, Lincoln Labs, and others, it allowed us to start becoming predictive on where we thought the virus might be two weeks from now, and then overlay that on where might we have capacity shortfalls in the nation.”
Bottom line, if you want to know what areas of the country might be hard hit two weeks into the future, the military is a step ahead of you.
JADC2 isn’t a finished effort; it’s only just begun. In the years ahead, the Air Force, working with the Army, Navy, Marines, Coast Guard, and Space Force will take the lessons from Stamford and try to turn those into a model for networked warfare: a massive web connecting every satellite, sensor, ship, jet, drone, and gun on the battlefield. The hope is that JADC2 will allow commanders, guided by artificial intelligence, to rapidly swap in one weapon or tool for another in the middle of war, presenting enemies with what military strategists call “multiple dilemmas.” In essence, if every weapon on the field can be perfectly controlled simultaneously, then the enemy can’t prioritize defensive measures; they don’t know which planes, guns, ships, to target first.
The COVID-19 response shows that data can move between the Air Force, Navy, and Army in a way that does allow everyone to know what’s going on. It’s been a dress rehearsal in a way that some military exercises aren’t.
“The forces that we have deployed across the nation for COVID support is very much a joint force: Army, Air Force, Navy that is deployed forward into each one of those areas, and so the datasets that we're working with from the beginning are also joint datasets,” said Strohmeyer.
The military sees JADC2 as essential to deterring Russia and China. In simulated wargames lately, the U.S. military isn’t faring well against those two potential adversaries. Pentagon strategists believe networked warfare will turn the tables.
“Achieving superiority against an adversary is all about making decisions faster than them,” Strohmeyer said. “The way that we make decisions is by getting the most accurate and aggregated data to that decision maker as quickly as possible so that they can achieve understanding of what's going on, then make a decision.”
NORTHCOM’s Stamford success only goes so far. There’s a big difference between getting a heads-up in time to send a few dozen doctors around the country, and employing that system to direct forces in combat against a technologically advanced adversary.
But “I’m more encouraged the faster I see them move to faster adoption of commercial technology,” said the ESRI official, describing the military’s JADC2 ambitions. “Now they’re making the agility the priority.”
In addition to lots of spending, experimentation, and purchasing of new software and other equipment, achieving the networked-warfare vision will also require commanders to learn to be more receptive to what predictive analytics algos are telling them. “This is something the commander is really driving towards, ultimately a digital transformation of the NORAD and NORTHCOM Combatant Command, but it's a digital transformation that isn't just doing tech differently, doing data differently. It's a culture change, and that's one of the biggest challenges that we've found is transforming our culture to think differently about that,” said Strohmeyer.
Getting top commanders to second-guess their own presumptions — their "gut” — and go with what the data is telling them is a big ask. But the pandemic has proven that a lot of presumptions won’t work for the future.