Enable breadcrumbs token at /includes/pageheader.html.twig

Recomposable Architectures: The Key to JADC2 Success: Sponsored Content

To successfully overmatch near-peer adversaries in the 21st century, the U.S. military requires decision advantage. Multidomain operations coordinate and bring to bear assets across all five domains of land, air, sea, space, and cyberspace.

To successfully overmatch near-peer adversaries in the 21st century, the U.S. military requires decision advantage. Multidomain operations coordinate and bring to bear assets across all five domains of land, air, sea, space, and cyberspace. Information dominance—getting the right information from the right sensors or systems to the right decision makers at the right time—is the key to victory on the multidomain battlefield of the future.

Joint All-Domain Command and Control, JADC2, is the path the Department of Defense has mapped out to achieve decision advantage.

It’s an ambitious vision, and a key leader tasked with achieving it, Lt Gen. Dennis Crall, USMC, the J-6 for the Joint Chiefs of Staff, doesn’t sound too impressed with industry offerings so far. “I am looking for good enough and I have not found a good enough,” he told AFCEA Hawaii’s Indo-Pacific TechNet virtual event in March.

Technology companies like Red Hat have built globally distributed, agile, reliable and durable architectures across government, healthcare, financial and telecommunication sectors. But the JADC2 use case presents some special challenges.

Traditionally, command and control of joint forces has been stovepiped by service—Air Operations Centers deploy air power in pursuit of joint battle plans, for example—but on the 21st century battlefield, operations must be executed across domains. Aircraft providing close air support one moment might be required to run a surveillance mission the next—a seamless transition managed at the tactical edge, without the need for time- and bandwidth-consuming command stovepipes to the rear echelon.

Victory relies on getting the right information from the right sensors to the right decision makers at the right time, so they can issue the right commands to the right platform. But the definition of “right” is different in each imaginable conflict scenario. The sensor-decider-shooter kill chain that JADC2 must service will be different in every engagement.

“The challenge JADC2 presents isn’t that every domain is different—there are obviously differences, but there are also important commonalities—the real challenge is that every engagement is different,” explains Phil Osip, principal solutions architect for Red Hat Public Sector.

Each engagement is going to apply force from one domain or set of domains against one or more others in a different configuration, using a differently structured kill chain. “If we have one way of doing things,” Osip says, “based on a particular scenario, we are going to be really, really successful in that one scenario, and really unsuccessful in every other scenario.”

Osip calls this “the ‘One Ring to rule them all’ fallacy”—the idea that one platform, one centralized infrastructure or one standard is going to provide the single solution to the multiple challenges of JADC2.

“You can’t have a hub-and-spoke construct against a near-peer adversary. If you have a central node, that is a big fat target. ... If you have a head, the enemy is going to cut it off,” he points out.

Indeed, as one critic has observed, near-peer adversaries like China have made a close study of how reliant U.S. forces are on “fragile long-haul communications networks and are developing concepts of operations that explicitly attack that vulnerability as a key initial task.”

To achieve resilience against technically capable near-peer adversaries requires a distributed, decentralized approach, Osip says. “The enemy can take out this piece or that piece, but those blows don’t have an impact outside their own blast radius. They can be absorbed.”

Decentralization enables resilient command and control in a contested environment. “Everyone has a plan until they get punched in the face,” observes Osip. “You don’t want to have a recognizable face that the enemy can punch.”

Decentralization also enables dynamic flexibility in capabilities. “There is no uniformity at an engagement level, so there ought not be uniformity in how our IT systems facilitate those engagements,” says Osip.

“In one engagement, I need to be the Millennium Falcon. In another, I might need to be a TIE Fighter. I can show up with a huge chest of toys, and hope I have one that’s right for every engagement, or I can show up with a box of LEGOs and dynamically compose what I need each time in an automated and repeatable fashion,” he says.

In IT terms, the box of LEGOs is a recomposable architecture—a dynamic and modular approach to IT provisioning that bridges both real and virtual infrastructure. Red Hat’s recomposable architecture has four characteristics that make it ideal for JADC2.

• Edge and Open hybrid cloud provides flexibility in infrastructure and provisioning.

• Microservices architecture provides scalability in capabilities and mission outcomes.

• Distributed integration and distributed data fabric provides secure data interoperability, without requiring adherence
to a single standard.

• Automation provides alacrity and repeatability of provisioning, policy/compliance enforcement, updating and auditing mechanisms with reduced timelines and external observability.

An open hybrid cloud strategy provides unified and seamless management of a heterogeneous IT environment combining multiple public and private clouds with on-prem infrastructure. It is essential for the military, even more so than for the complex global enterprises it was originally used by. In a shooting war, “You really can’t afford to leave all your eggs in one basket. And you also can’t afford that basket to be so very far away from you that you incur latency or a disconnect every time you want to try and use the eggs that are in it,” says Osip.

Open hybrid cloud also allows some uniformity across very different roles and situations for users, says Osip. “You don’t want to have everything be a snowflake. You don’t want to have the ubiquitous big cloud experience, be very, very different in how I consume and produce [data] than it would be at the division level or the battalion level or even the squad level. You do want some uniformity and an open hybrid cloud strategy provides you that.”

Red Hat’s edge and open hybrid cloud enables a distributed and dynamic infrastructure that supports modern cloud-native application architectures while permitting existing monolithic application architectures to garner the hybrid cloud benefits across echelons. A modern microservice and container-based architecture enables scalability and dynamic deployment of new capabilities. In practice, Osip says, that means “in the tactical environment, I don’t have to add new hardware to be able to burst consume [IT] resources for periods of time. ... That’s a really fundamental shift.”

Finally, Red Hat’s recomposable architecture enables a data-centric approach through distributed integration. “Lt. Gen. Crall said it,” points out Osip, “‘It’s all about the data.’”

But Crall has been misunderstood, he adds. “That doesn’t mean JADC2 is a data storage challenge. It doesn’t mean it’s a data transport challenge. ... The challenge here is about understanding data modalities and topology: What data needs to be at rest, stored, right now and where—that’s important because you might need it—what data needs to be in use and where; what needs to be moved, and where and how fast. ... Understanding all that is where the seven Vs come in.”

There are various versions of the seven Vs. This is Osip’s:

• Volume

• Velocity- The classic three Vs of old style big-data. “Pretty self explanatory.”

• Variety

• Volatility- Data has a shelf life. “The best data in the world is useless unless you can act on it in time,” says Osip. “In JADC2, which is all about speed to decision, data volatility tends to increase as you get closer to the tactical edge.” There are plenty of uses for historic datasets further to the rear—training artificial intelligence and machine learning models for instance. But when boots are on the ground, seconds count.

• Value- “Value is multidimensional. The value of data at rest is going to tell you, is this something that I need to move to backup storage? Or do I need to keep it live in primetime? If it’s data in use, the value is going to tell you, can I introduce an efficiency by offloading some of the data as metadata, so that I’m focusing on
a subset of it for that data in use.”

• Veracity- “To ascertain the quality and accuracy of data, I may need to know other things, contextual things. What’s its provenance? Where does it come from? Who collected it?” Addressing and understanding the inconsistencies and uncertainty of data is essential to scoping and scoring its utility

• Validity- Even true information isn’t always useful, if it’s irrelevant, or out of date, or in the wrong place.

Distributed integration enables JADC2 to bring the seven Vs to bear on the battlefield. “There are sensors all over the battlefield. The first O of the OODA [Observe, Orient, Decide, Act] Loop is already distributed. Distributed integration is about securely distributing the other three functions, too,” explains Osip.

Automation provides the ability to deploy, maintain and extend all of these capabilities in a secure and repeatable manner while compressing the OODA loop timeline.

Data has to be worked on and transformed into actionable intelligence. Over time, intelligence builds a corpus of knowledge, often generating insights and the possibility of accumulated forensic wisdom. “Historically, that data journey, from raw data to intelligence to knowledge to insight and finally wisdom, that happened as you went up the echelons, through the chain of command. Only at the top did you have the whole picture. But for multidomain operations, for JADC2, that journey has to happen at or much closer to the tactical edge. All the different phases have to be brought to bear there. That’s how you distribute and compress the OODA loop. … That’s how you win.”

For more information visit redhat.com/dod.