UPSKILLING: Blaine Perry, Redstone Test Center chief data officer, leads a data science upskilling class for members of the Redstone Test Center digital transformation community in June 2024, at Redstone Test Center, Alabama. (Photo provided by Maj. Lucas Gebhart)
Data mesh revolutionizes U.S. Army Test and Evaluation Command’s ability to help speed capability to the hands of warfighters.
by Maj. Lucas Gebhart and Blaine Perry
In 2015, the scourge of ISIS swept Syria and the western deserts of Iraq, tearing apart the hard-fought gains from years of Americans fighting in Iraq. The resulting operational environment was confusing and dangerous, with myriad factions of friendly and enemy forces in a nonlinear battlefield complete with widely dispersed high-level threats. Not since the opening days of Operation Iraqi Freedom, before the development of the Common Missile Warning System (CMWS), had Army aviation encountered such a nonpermissive threat environment. Air defense artillery assets—like 57 mm anti-aircraft cannons and the high-rise building structure of Mosul—drove rotary-wing aviation to increased altitudes, while other threats with questionable CMWS effectiveness drove the need to remain low level.
Given the currently fielded aircraft survivability equipment on many rotary-wing airframes, the threat environment was not conducive for safe and effective deployment of the AH-64 aircraft that the Iraqi army so desperately needed over its shoulders to defeat ISIS. U.S. Army aviators needed access to a new aircraft survivability equipment capability. This unforeseen requirement drove a Joint Urgent Operational Needs Statement (JUONS) for fielding of a new critical capability, requiring the expertise of the testing and evaluation professionals at U.S. Army Test and Evaluation Command (ATEC) to conduct a thorough but expedited evaluation.
PLAN AND EXECUTE
ATEC, in direct support of Army Futures Command, is expected to plan and execute rigorous testing of the systems the Army intends to field, gather large amounts of data from those systems, reduce that data into meaningful insights and share the findings with senior leaders. In the general case, these tests can span multiple years of effort, and thousands of hours of labor from the test community over various iterations of testing and data analysis.
JUONS efforts are unique. Though ATEC is still required to plan and execute testing and gather data for reduction into insights, the needs of the community demand that this process be condensed to months versus years. This urgency typically requires 18-hour days and seven-day weeks. Teams would start their days at 4:00 a.m., plan, brief and fly several sorties (planned flights with specific routes and data collection objectives) and hope to begin data reduction at 5:00 p.m.
One line of testing, known as bias and jitter, characterizes the performance of a pointer-tracker on the aircraft in terms of operating power and laser movement on target. It is complex, involving thousands of frames of infrared imagery, aircraft time-space-position information, video processing, and data sizes and complexity exceeding the processing capability of ATEC’s tool set.
To make things more difficult, mission requirements dictated testing at multiple test sites across the U.S., most often in areas with limited or no connectivity, meaning that data delivery involved shipping hard drives back and forth across the country. Not only could ATEC not conduct timely processing of data, but it also took weeks to get the data to analysts and evaluators, resulting in months from test event to informed decision. At a time when the Army and ATEC needed an efficient observe, orient, decide, act (OODA) loop, legacy systems failed the evaluators doing everything they could to get capability to the field.
To reduce this cycle and meet the needs of the Army, ATEC needed a more efficient means to transport data to analysts, better systems, tools and processes for analysts to work with data sets rapidly growing in size and complexity and serve evaluations and insights to the acquisition community.
As ATEC struggled through the JUONS cycle, 4-6 Heavy Attack Reconnaissance Squadron (4-6 HARS), 16th Combat Aviation Brigade was finalizing a movement from Fort Carson, Colorado to Joint Base Lewis-McChord, Washington, having deferred many long-term and time-intensive maintenance tasks until after the move. Not yet slated for deployment, but imbued with a healthy dose of paranoia, one of the authors, then-Capt. Lucas Gebhart, developed a plan to bring the organization back to a healthy deployable state by synchronizing two to three years of maintenance with a busy training schedule in a 12-month period.
Over the course of that year, 4-6 HARS would accomplish that maintenance, learn of, prepare for and initiate deployment to Iraq and learn of, train on and field the JUONS requirement. This was disruptive to the unit’s deployment training plan, requiring tens of hands-on hours per airframe over weeks of downtime per airframe in addition to the already robust maintenance requirements in a unit trying to train and deploy to Iraq. More importantly, from the perspective of the observing aviators, they were giving up a security blanket with over a decade of proven results (the CMWS) in exchange for a system that seemed buggy, hastily fielded and with which they flew precisely zero hours prior to taking off for the first time in a highly contested environment. To put icing on the cake, in addition to the training and maintenance disruptions from JUONS installation, fielding disrupted block leave afforded to Soldiers pending a yearlong rotation to Iraq.
A PATRIOTIC PERSPECTIVE: Then Capt. Lucas Gebhart, left, and 1st Lt. Matthew DiPinto stand with the American Flag in Erbil, Iraq, in January 2017. Gebhart helped his organization synchronize several years of maintenance with a busy training schedule. (Photo provided by Maj. Lucas Gebhart, ATEC)
AN INFLECTION POINT
In early 2022, recognizing the shortcomings of its data analysis and delivery capability, ATEC hired a chief data officer, with the mandate to digitally transform the way that ATEC conducts the core of its mission. Bringing together a group of experts, and leveraging environment and technical assistance from U.S. Army Futures Command, ATEC set out to design and implement the ATEC Data Mesh. A data mesh collates data along logical business lines, while maintaining access control at the data owner level, leveraging robust governance and cataloguing to enable federated search across the full data corpus. By November 2023, ATEC brought the data mesh from ideation to authority to operate, the core requirement for use of an information system in the DOD, allowing ATEC to onboard its user base and begin to transform data use across the command.
| WHAT’S A DATA MESH?
A data mesh makes data more readily available to stakeholders by storing data in a common location with separation along logical business lines, while enabling data owners to control access. The ATEC Data Mesh breaks data down into data domains based on test mission areas, appends appropriate metadata and catalogues data at upload to enable federated search. When fully implemented, stakeholders will no longer need to scour discs, hard drives or unformatted file architectures to find data for a system under test—they will be able to find all related data via a single search utility. |
The ATEC Data Mesh seeks to resolve both technical and bureaucratic hurdles to efficient operational conduct, with the ATEC Cloud Enablement Team providing the infrastructure, technical support, funding and managed services to the entirety of ATEC. The construction includes continuous integration, continuous deployment pipelines to emulate a robust commercial environment within the security required by DOD. Taking the “Field of Dreams” approach, designing and delivering world-class capability for any interested party in ATEC, the cloud team built and deployed data domains encapsulated in ATEC’s authority to operate for each of ATEC’s subordinate elements. After onboarding, users will find a robust set of database servers (e.g., SQL, SMI, Mongo, PostGres), data analytics tools (e.g., Databricks, Synapse, Function Apps) and an Azure Virtual Desktop pool (a virtualized computer infrastructure) complete with the software tools that ATEC’s professionals use on a day-to-day basis (e.g., Python, R, JMP, MatLab). The cloud team’s objective is to enable any analyst, any data, anywhere.
The core of the ATEC Data Mesh is the common reference model (CRM), an application program interface (API) infrastructure that seeks to link to root authoritative sources of data throughout the Army, while exposing the root authoritative data for test and evaluation. The CRM will enable reuse of this root data in applications and software within ATEC and the Army, allowing developers to form applications without risk of the underlying data becoming stale or incorrect. ATEC’s portion of the CRM is the ATEC Enterprise Framework, a systems modeling language model of the entire ATEC enterprise, its capabilities and its business processes. Expanding on this work, ATEC is building an evaluation framework that seeks to standardize measures of performance and effectiveness used to evaluate a system’s performance against the root requirements document—the source of truth for what constitutes “passing the test.” In the end, the CRM, the ATEC Evaluation Framework and the ATEC Data Mesh will enable ATEC to digitally link an acquisitions program all the way from the requirements document to the test data and evaluation of each requirement. With all of the test data for a program residing in their common home in the ATEC Data Mesh, digitally linked via the CRM, ATEC can drastically improve both the efficiency and effectiveness of ATEC’s test and evaluation mission.
In less than two years since ideation, ATEC developed and deployed a capability to its workforce and finds itself in the nascency of adoption. While there is much to do in terms of executing the vision of the future outlined above, the ATEC Data Mesh is currently capable of executing its any analyst, any data, anywhere mission.
MAKING A LIST, CHECKING IT TWICE: Redstone Test Center Civilians Chance Graham, outside rear, Blaine Perry, outside front), Charles Packard, back seat, and Patrick Atkins, front seat, conducting installation checks in preparation for a JUONS flight test at Redstone Test Center, Alabama, in January 2016. (Photo provided by Blaine Perry, Redstone Test Center)
ATEC’S FUTURE: DATA AS A WEAPON
Because of the ATEC Data Mesh, the ATEC of today and tomorrow is no longer restricted to the systems and processes that have inhibited our ability to deliver capability at the pace of need. In the past, because of the timing and locations of testing, we struggled to access the computing resources or data storage that we required. Today, we have the full capability of a Microsoft Azure cloud environment and have worked to bring together teams of government and commercial technical experts to revolutionize the way that ATEC does business.
To highlight those capabilities and show the ATEC community that very complex problems can be solved using the new tools available, ATEC’s chief data officers initiated several data pilots, including development on the bias and jitter problem. Redstone Test Center created an application in the ATEC Data Mesh that would automatically ingest raw video files, locate the laser on the target board and calculate its position error. In the past, each two-hour sortie required an average of 10 hours to curate files, trim video and extract information—a process that analysts could not start until the data arrived. Over the course of the JUONS effort, ATEC completed dozens of sorties. Today, after three weeks of effort by Redstone Test Center’s data scientists, a user can load an entire sortie’s video data directly to the ATEC Data Mesh from the field site via our data upload utility HERMES. The application will automatically identify files, conduct bias and jitter processing, and produce results in minutes. Importantly, that data can be shared with all ATEC’s partners instantly from anywhere as soon as the automated processing is complete. The ATEC Data Mesh can decrease the turn time of the bias and jitter testing from months to hours. Relevant, timely information to senior Army leaders is a core part of achieving ATEC’s mission. The ATEC Data Mesh has redefined these terms, enabling us to provide more relevant data at speeds that were never possible.
This is not the end for the ATEC Data Mesh: ATEC’s data is a weapon and a force multiplier. In the case of the JUONS, just one additional month to view the JUONS test data and to ask questions would have meant so much to the confidence of a Soldier being sent into harm’s way, entrusting their lives in a system with which they have never flown. Extending this problem set to weapons and munitions, master gunners and technical experts in formations would make use of trustworthy and accessible data to expand the lethality of their fighting formations.
ANY QUESTIONS?: Maj. Lucas Gebhart, ATEC deputy chief data officer, leads a white boarding session on the CRM at Aberdeen Proving Ground, Maryland in May 2024. (Photo provided by Maj. Lucas Gebhart, ATEC)
CONCLUSION
The future of warfare relies more heavily on data than ever before. Artificial intelligence, machine learning, and more, all require enormous amounts of high-quality data—the kind of data that the ATEC community produces today. In the past, Gebhart would not have had access to this data, but with the ATEC Data Mesh, he can. In turn, ATEC can make this data descriptive, predictive and prescriptive, telling Soldiers not just what happened in a test, but describing what might happen in other environments, and how to gain an advantage by using the systems to their maximum extent. By reducing its OODA loop, ATEC can enable Soldiers in the field to reduce theirs.
For more information, contact Maj. Lucas Gebhart at lucas.c.gebhart.mil@army.mil.
MAJ. LUCAS GEBHART is the ATEC deputy chief data and analytics officer. He holds an MBA from Harvard Business School and a B.S. in operations research from the U.S. Military Academy at West Point. During his career, Gebhart has deployed four times to Iraq and Afghanistan, most recently as a troop commander of the AH-64E Apache troop supporting the Iraqi army in regaining control of the city of Mosul in 2016-2017.
BLAINE PERRY is the chief data and analytics officer at the U.S. Army Redstone Test Center. He holds an M.S. in business intelligence and data analytics from Carnegie Mellon University and a B.S. in aerospace engineering from the University of Alabama. In 2020, he was selected as a member of the inaugural cohort of the U.S. Army Artificial Intelligence Scholars program, where he spent two years with the U.S. Army Artificial Intelligence Integration Center in Pittsburgh learning how best to leverage artificial intelligence and machine learning within the U.S. Army.
