Repairing aircraft such as the Lockheed Martin F-35 Lightning II fighter jet and the Boeing B-52 Stratofortress bomber can take months due to a lack of data interoperability between the necessary repair parties.
Aircraft maintenance officers take apart engines, inspect them and send data back and forth to the Air Force Research Laboratory (AFRL) and the Air Force Materiel Command’s (AFMC) Air Logistics Centers (ALCs) to determine the best repair approach.
The Defense Department’s newest software factory, Hangar 18, in January organized at the Wright-Patterson Air Force Base in Dayton, Ohio to solve this data interoperability challenge, accelerate the rate of repairs and research new, better ways to manage mission-critical data.
“[It] results in months — months of time when you're taking apart an engine and inspecting it and sending the data back to the AFRL and the Air Logistics [Center] so it can get concurrence on the best repair approach for this particular piece or component,” Hangar 18 Director Matt Jacobsen said in an interview with GovCIO Media & Research. “It's a mess.”
As the Defense Department becomes more data- and software-dependent, software factories such as Hangar 18 become more critical to the future of the department’s IT infrastructure.
Software Factories Get Elevated at DOD
Air Force Secretary Frank Kendall said software factories do “certain things” very well and wants to “tailor” software factories to bring the most value to the Air Force.
“The two places where I think [software factories] make sense [are] the Air Operations Centers — Kessel Run has been doing a lot of work to incrementally upgrade those capabilities and that’s a reasonable thing for a software factory to do — and sustainment over time. We have good examples for that with our weapons systems,” he said at the Potomac Officers Club Air Force Summit Tuesday.
Air Force CIO Lauren Knausenberger said there’s no one true definition of a software factory due to its flexible nature, but believes they will collectively help transform Air Force IT.
“Sometimes ‘software factory’ means we’ve built a pipeline and supply developers with tools," Knausenberger said during a fireside chat at the Air Force Summit. "Sometimes it is developers solving a problem or airmen coming together and problem-solving and doing design-thinking, user-centered design and outsourcing the development."
Knausenberger said she wants airmen to solve software problems “on the fly,” signaling a desire to invest heavily in cybersecurity and IT and “execute the heck out of it.” She described zero trust, identity solutions and hybrid cloud as “game-changing capabilities” to that end.
“To get that speed to warfighter, we have to have that cloud everywhere in the world and have that software on demand,” she added.
Poor Data Management Hurts the Mission
Hangar 18 — named for the legendary hangar at the Wright-Patterson Air Force Base said to be a repository for UFO evidence and research — focuses on enabling secure data exchange between disparate IT systems such as the Air Force Life Cycle Management Center, which is managed by the 88th Air Base Wing at Wright-Patterson.
The Air Force struggles to manage lifecycle data, and sometimes crucial datasets are siloed across different teams and organizations within the department, Jacobsen noted.
“We had this interesting pattern that had emerged over the last several years starting in 2018 — my background is in research data acquisition research exchange and categorization, so I build data research systems,” Jacobsen said. “What started happening was several years ago, we were approached by a couple different programs that were not as research-oriented and working on manufacturing capability of critical systems.”
Research around new ways to manage data, especially with emerging technologies like artificial intelligence, is flourishing, but researchers often struggle to translate their data and software research efforts to the real world.
DOD’s cybersecurity protocols present additional hurdles to data sharing.
“Processes are locked up behind firewalls, and there's no process of easily moving that data,” Jacobsen said.
The inability to share critical data efficiently, effectively and securely has real-world consequences on the rate of repair for critical aircraft such as the F-35 and B-52.
Hangar 18 hopes to solve that problem with cloud technologies, data research and DevSecOps.
“What Hangar 18 has done is build cloud infrastructure and data tools to allow us to do this type of data exchange in minutes,” Jacobsen said. “We've been commissioned to build a data hub for AFRL. It is the data integration platform for the whole lab, so we're talking hundreds of connected systems. It’s an enormous project that will house data of all forms. It will allow us to ask complex questions about the type of work that we do.”
The goal of the data hub is to accelerate research to be more efficient and integrate into broader efforts for data engineering, he added.
Jacobsen also said zero trust “absolutely plays a role” in accelerating data exchange because it establishes a “chain of custody” of the data. AI holds promise, but despite “hotspots of innovation” at organizations such as the Defense Innovation Unit (DIU), Jacobsen thinks DOD’s data management isn’t ready for AI transformation.
“We have a program manager (PM) that comes to Hangar 18 saying, hey I need help managing my data, and we'll say, what do you need, and he'll say, gather it all — and that's not a great strategy,” Jacobsen said. “A lot of these PMs are not literate in data practice, so we don't have that relationship. We are in the trenches. We have a very long way to go before we see widespread use of AI and ML in this space.”
Value of a Flexible Hiring Approach
Lacking a skilled workforce is another challenge as DOD continues to digitally transform. Holding a single degree in computer science or mechanical engineering is no longer enough. Jacobsen said he needs workers with both sets of skills.
“We’ve seen the emergence of a hybrid practitioner, a data engineer or wrangler, who has a strong background in the subject area who also are very strong programmers,” Jacobsen said. “These work hand in hand with practitioners to see how the data all relates.”
But there aren’t enough workers with both sets of skills, which is why workforce training and upskilling are major priorities for Hangar 18.
“The third bullet on our playbook is to train the workforce,” Jacobsen said. “We try to avoid the workforce-for-hire model. We prefer to see this as a partnership.”
Jacobsen’s comments point to the trend of DOD increasingly blending government and industry to quickly develop smarter solutions for evolving, mission-critical military needs.
For example, techniques used by software factories such as model-based engineering and digital twin design — which is a virtualization of a program or system that spans the lifecycle of the product to allow for improved fine-tuning of the end product — come from industry.
Kendall said he wants the Air Force to continue partnering with industry to develop software solutions.
“The philosophy of [model-based engineering and digital twin] design has proliferated over time, and I think this is a model for the future, program offices living on the same digital platform as the contractors doing the design … and they work together to develop functionality. I think that’s going to be the norm going forward,” he said at the Air Force Summit.
Kendall’s comments align with Knausenberger’s and DOD’s new software modernization strategy released in February, which hails software factories as the model for DOD software development going forward.
To that end, Jacobsen called on Air Force leadership for more buy-in so software factories and other “on the ground” innovators can pursue new, potentially game-changing solutions for the Air Force.
“My plea for leadership is to continue offering pathways for innovation and empowering lightning-fast innovation,” Jacobsen said. “Digital transformation is moving froward in a lot of years, but it's still very high-dollar, large-scale efforts. Having support to move quickly and do rapid prototyping in cloud environments … improving the ability of software factories and citizen developers to have access to robust cloud resources, giving them a more diverse set of options moving forward is going to be critical. There's a lot of uncertainty in this space. A lot of people aren't even aware of what’s out there.”