2

Workshop One, Part Two

OPENING REMARKS

Workshop series co-chair Lt. Gen. Michael Hamel (USAF, ret.), independent consultant, welcomed participants to the second day of the workshop series. Workshop One chair Gen. James (Mike) Holmes (USAF, ret.), senior advisor, The Roosevelt Group, recapped the previous day’s discussion on Air Force strategy and needs in preparation for the forthcoming panel presentations from the capability providers and data contributors pursuing those challenges. Workshop series co-chair Ms. Deborah Westphal, chairman of the board, Toffler Associates, encouraged participants to share their insights as to whether the Department of the Air Force (DAF) is on the right path toward digital transformation. Mr. Edward Drolet, deputy director, Chief of Staff of the Air Force Strategic Studies Group, noted that the restrictions of the COVID-19 pandemic motivated some of the DAF’s digital transformation progress, and he emphasized the value of learning lessons during a crisis. In light of recent operations in Afghanistan, he expressed his hope that access to information for commanders to make decisions would continue to increase.

Gen. Holmes shared an anecdote about Air Combat Command’s (ACC’s) responsibility to coordinate rescue operations after a hurricane several years ago. He explained that because no effective Air Force tool was available for this task, airmen created a common app for rescue operations within hours using a chat app and Google maps. Airmen are compelled to operate outside of the Department of Defense Information Networks (DoDIN) and use their own coding knowledge to create one-off tools because DoDIN does not meet the airmen’s needs. He suggested that the DAF leverage existing technologies and tools that make it easier to share information instead of taking several years to build its own capabilities. However, he cautioned that the Space Force, given its increased efforts to provide coding education, could have challenges with version control; the likelihood of vulnerabilities increases as modifications are continually made to software. In response to Gen. Holmes’s concern about version control, Dr. Paul Nielsen (USAF,

ret.), director and chief executive officer, Software Engineering Institute, Carnegie Mellon University, said that the commercial world has faced and overcome this challenge.¹

Reflecting on the previous day’s discussion, Ms. Westphal questioned the DAF’s 30-year budgeting outlook, given how difficult it is to predict the future state. Gen. Holmes suggested focusing programming on a 5- to 7-year time frame instead so as to achieve better results. Dr. Marv Langston (USN, ret.), independent consultant, added that by 2027, there could be 12,000 low-earth orbit Starlink satellites floating around the earth, and denied, disrupted, intermittent, and limited bandwidth problems will be unlikely for those outside of the Department of Defense (DoD) system. Dr. Nielsen explained that the DAF’s hurdle is not simply a technical problem that can be solved with better infrastructure and faster decision making. Instead, there is a human element, for which practice and exercise are critical. He emphasized that the improvement process for the DAF will be incremental.

LEADERSHIP FOR TRANSFORMATION

Mr. Anthony Reardon, Senior Executive Service, Administrative Assistant to the Secretary of the Air Force, explained that the DAF lacks a cogent strategy for digital transformation—without a unity of effort, disparate solutions arise among the several Air Force chiefs. Another challenge is resource management. He underscored the need for the DAF to become more competitive with industry, by hiring individuals who have experience with transformation on a broad scale.

Mr. Reardon described Disney World as a model of the development of integrated systems. Over the course of only 1 year and $1 billion, the theme parks successfully united their data and operations. Despite the fact that 4 of the theme parks were each separated by a decade and each built on a different IT backbone, Disney World integrated 7 theme parks, 27 resorts, 100 attractions, and 150 restaurants, along with its reservation system. The DAF struggles to implement a similar transition on a larger scale—for example, the DAF is located in several countries and has a multitude of systems, some of which are 50 years old. He emphasized that because the DAF’s budgeting cycle is too short to build an integrated plan for the future, people find their own near-term solutions that are not interoperable. For example, while everyone supports the development of a program that interfaces with the Mission Partner Environment, no one wants to work toward the standards that were established by the program.

Mr. Rich Lombardi, Chief Management Officer, Office of the Under Secretary of the Air Force, reiterated that the first step in any transformation is the development of a coherent strategy. Although many across the Air Force are doing exceptional work, individual units remain unintegrated in stovepipes and have to stitch activities together to complete missions. He observed that although people claim to support an enterprise system, they remain reluctant to collaborate. Once a strategy is developed, he continued, a governance process will be needed to synchronize efforts among people, processes, and technology. Furthermore, achieving an enterprise system is not possible without enterprise-level funding, and there is a balance needed so that this enterprise-level funding is protected and functional areas retain flexibility. Another problem is that everyone wants to develop their own case management systems, yet there is no need for 20 different case management systems on the air staff, especially given that most likely seek the same data and none are fully funded. Instead, because the requirements are likely 90 percent similar, he suggested pooling the funding—his team is building one such unified case management system as a pilot. Pilots are also under way with the AI community to build a concept of operations (CONOPS). The ultimate goal is to eliminate many manual tasks so that airmen have more time to do their jobs. A service catalogue of capabilities is also being developed, which could be used by functional leaders to better determine gaps. Mr. Lombardi’s team has partnered with a program executive office (PEO) for business systems at Maxwell-Gunter Annex Air Force Base to begin that effort. Although progress continues to be made in siloes, the objective to move toward an enterprise perspective remains. In closing, he remarked that it is

___________________

¹ Commercial software companies have developed and adopted automated version control and collaboration tools that address this problem. DoD contractors and organic coders need to adopt these tools.

critical to ensure that the innovative work of the airmen impacts the overarching desired capabilities of the DAF.

Open Discussion

Lt. Gen. Hamel wondered if there is consensus across the DAF on the need for a digital strategy and, if so, how it would be created. He also asked how to harmonize roles, responsibilities, resources, and governance across the DAF. Mr. Reardon acknowledged that while most understand the need for a cohesive and coherent strategy, many do not understand the specific purposes of programming—there is no strategy that unites different components and assigns responsibility for each. The most significant barrier, however, is the translation into a resource requirement. Part of the challenge, Mr. Lombardi added, is that “digital” has many definitions. The chief software officer, chief architect, chief technology officer, chief data officer, and chief information officer play essential roles, but it is important that all of their initiatives connect. He explained that the current model of governance is not set up to govern a department of two services; a new model of governance is needed in which only optimal innovations based on enterprise goals and mission needs are funded.

Recognizing the vastness of the future digital environment of operations, Ms. Westphal asked if a CONOPS and doctrine are driving the DAF’s strategy. Although he was not aware of any CONOPS driving the strategy, Mr. Lombardi acknowledged the urgent need to understand the facets of “digital” to determine how to develop the strategy. Ms. Westphal added that a misplaced emphasis on what the DAF will purchase still exists. Gen. Holmes noted that the A5 has been tasked to develop an overall vision of the future environment and the threat, including the CONOPS that will drive requirements and platforms. However, he wondered whether the “underlying digital backbone” is receiving adequate attention. Mr. Reardon explained that there are many steps that need to taken before the joint all-domain command and control (JADC2) environment that is being constructed could function effectively. Capabilities are being built that may not integrate seamlessly into JADC2 or future constructs, simply because money is available. He asserted that a strategy for incrementally building capabilities would be more beneficial.

Dr. Julie Ryan, chief executive officer, Wyndrose Technical Group, discouraged “digitization for digitization’s sake.” Instead, processes could be reengineered and adapted to technologies to accomplish goals, and become more efficient and effective. When processes are reengineered, they have to be reengineered with consideration for modern threats. A critical area that has been systematically ignored by program developers thus far is security—confidentiality, availability, integrity, protection, detection, and correction. “Bolting on” instead of “baking in” security is ineffective, she continued. Mr. Reardon commented on the value of zero-trust architectures and noted that the DAF is working to integrate security from the beginning. Mr. Lombardi remarked that when people become enamored with technology, they lose sight of the difficult work required for process engineering and try to modify the technology to fit existing inefficient processes. He suggested reviewing processes and determining which technology could help achieve the goals instead of allowing the technology to drive the solution. Mr. Reardon added that outcomes have to be the focus when reengineering processes. Dr. Annie Green, data governance specialist, George Mason University, said that it is impossible to reengineer something that has not been engineered. And if something has not been engineered, she continued, it cannot be managed. She emphasized the need to focus not only on digital technology but also on digital representation. When the components of an enterprise are represented in digital form and better understood, they can be used for both strategy and operation. Another element missing from the current framework is the levels of abstraction in the enterprise in terms of strategic, tactical, operational, and cognitive decision making, as well as the intelligence. Instead of repeating what has been done in the past, she highlighted the value of understanding steps forward as current operations are maintained and improvements are being built for the future. Mr. Martin Akerman, Chief Data Strategist, Office of the Chief Data Officer, DAF, agreed with Dr. Green about the need to better use the knowledge layer as well as the data fabric that drives it. He described Air Force leadership as well positioned to change the culture and target the risk appetite of the DAF in the movement to a new model of

decision making. Dr. Green added that digital transformation also requires consideration for paradigm changes, and she suggested focusing on the integration of knowledge management in the flow of digital transformation.

Mr. Alden Munson, senior fellow and member, Board of Regents, Potomac Institute for Policy Studies, observed that the DAF tends to select commercial products that are the best approximation to its requirements and modifies them to better fit its needs. He advocated for another step in the process: after selecting the commercial product that has the best approximate fit, go back to the requirements and the architecture, and consider whether the requirements need to be modified. He emphasized that the standard “requirements, to design, to implementation approach” is ineffective; flexibility is needed based on the availability of commercial products. Dr. Ryan added that it is not feasible to avoid commercial products; furthermore, because many products are rushed to market, it is critical that contracts demand proven security before product delivery. She proposed maintaining products in a sandbox to avoid downstream liability with other parts of the architecture.

Gen. Holmes perceived the desired outcome of the DAF strategy to be collecting available data and using them to sense better, to make decisions better, to command and control better, and to support headquarters better; he wondered whether there are indicators to monitor those outcomes. Mr. Jay Santee (USAF, ret.), vice president, Strategic Space Operations, Defense Systems Group, The Aerospace Corporation, observed that process improvement should be based on empowering individuals to make faster and better data-driven decisions. He explained that requirements are driven by the decisions we want to make, the information we want available, and the person we want in charge of making that decision. Thus, a mechanism to pass requirements to developers to improve processes is needed. He asked about the DAF’s broader plan for process-driven, data-driven decision making by empowered individuals at a proper level of authority. Mr. Lombardi commented that over the past year, the DAF has begun to empower decision makers at different levels, but the DAF does not yet have a solid understanding of the connective threads of processes. In closing, Mr. Lombardi noted that the DAF is aware of its blind spots and the need to eliminate stovepipes, as well as the need to focus on how technology will integrate with current and future plans before acquiring, but it first needs to develop a coherent strategy to achieve digital transformation.

ORGANIZATION AND TECHNOLOGY FOR TRANSFORMATION

Ms. Lauren Knausenberger, Chief Information Officer, DAF, explained that consistent agility could lead to consistent advantage—a feature that the DAF currently lacks. She described four pillars to enable seamless JADC2 across the enterprise:

1. Creating a digital foundation, including reliance on a 21st century network, availability of cloud-based services, use of a DevSecOps platform, development of a zero-trust framework, integration of cybersecurity that protects against ever-changing threats and reduces overall threat surface, and creation of a data fabric.
2. Focusing on the user experience for warfighter effect by understanding what tools and data the warfighters need to be more effective in their missions, what barriers exist, how to measure and improve performance of tools with automation, and where to insert IT-driven capabilities.
3. Enabling digital talent, with initiatives such as Digital University,² retention incentives, and diversity and inclusion efforts. Digital University provides a path to upskill and/or reskill the entire existing force, maintain technical edge by equipping airmen with digital skills, and recruit talented individuals that can help move the DAF forward with its digital transformation.
4. Attacking manual process, outdated policy, and redundant IT. In Operation Flamethrower, for example, airmen advise which systems should be eliminated and which should be sustained, with the goal to open tradespace and better spend funds.

___________________

² The website for Digital University is https://digitalu.af.mil, accessed November 23, 2021.

On the journey of digital transformation, Ms. Knausenberger commented that first, digitally savvy airmen and guardians could help adapt technology and find new ways to conduct missions. Second, the right apps would enhance the fight of the future—for example, a future in which the majority of weapons systems are integrated software platforms. Third, digitization would be enabled by an artificial intelligence (AI)-ready technology stack at scale (i.e., AI, machine learning [ML], sensors/Internet of Things, data fabric, cloud/edge computing, and connectivity). With the addition of zero trust, the use of more reliable commercial lines could be a low-risk way to improve connectivity. She summarized the coordination of these DAF efforts as follows: airmen and guardians and their missions are at the core of the DAF. Enterprise IT delivers the data that are needed to operate in a contested battlespace. Data then feed into advanced applications, which require usable data (e.g., via the Visible, Accessible, Understandable, Linked and Trustworthy [VAULT] data platform; Core Data Service; Unified Data Library), and those data require a reliable network. Enterprise IT (i.e., the digital foundation, with services such as Cloud One³ and Cloud-Hosted Enterprise Services) capabilities support the warfighter and the mission. The data fabric is an essential part of how all of these steps connect.

Ms. Knausenberger explained that all of the DAF’s public-facing websites are on Cloud One. Cloud One’s recent statistics include 19 million monthly log-ins, and more than 600,000 prevented attacks in August 2021. The next objective is to move more of the secret enterprise into the cloud. Platform One, which allows the DAF to develop, deploy, operate, and sustain software in a secure and agile manner, can deploy code 49 times per day to a warfighter in locations permitted by the infrastructure. The goal is to increase code deployment to thousands of times per day in the future. Iron Bank, the Air Force’s container repository, has more than 700 products and services that can be used securely across any environment.

She emphasized that zero trust offers security, simplicity, and accessibility, and enables the Advanced Battle Management System and JADC2. Zero trust allows a transition from an environment of sophisticated cyber threats, diminishing warfare advantages, and constrained technology to an environment that imposes a cost to the adversary, offers the freedom to operate, and provides a cyber-enabled warfare advantage. Identity, credential, and access management and critical data tagging are foundational to zero trust (see Figure 2.1). While newer software is already built for a zero-trust environment, older software has to be brought into the future. With the underlying data fabric and zero-trust architecture, she expressed her hope to advance to one Secret Internet Protocol Router Network (SIPRNet) for warfighting and then ultimately to one collapsed multi-level classification environment, where it would be possible to collaborate across the joint community and with allies and to release software updates in real time.

Ms. Knausenberger described the Joint Warfighting Cloud Capability, which is an enterprise-wide multi-vendor cloud solution that could be available for use globally in June 2022. This capability could increase the speed of moving data to the tactical edge. However, she emphasized that the DAF has had more than a decade of underinvestment and has much work to do to move into the 21st century of digitization. Although she asserted that the focus should be on investment in modern infrastructure for digital transformation instead of in the sustainment of aging infrastructure, it is difficult to abandon existing technology without up-front investments for new technology.

Ms. Eileen Vidrine, Chief Data Officer, DAF, explained that her team works closely with the chief information officer’s team to move products forward and develop a common architecture across the DAF. The priorities of the chief data office (CDO) include operationalizing the seven strategic goals of the DoD data strategy: visible, accessible, understandable, linked, trustworthy, secure, and interoperable.

___________________

³ Cloud One is a multi-cloud environment that hosts the Air Force’s enterprise general purpose applications.

Ms. Vidrine noted that the first step to achieving this vision and improving efficiency is the democratization of data. The DAF is currently piloting a data catalogue that will make data more visible and accessible by federating to the DoD data catalogue. It is essential that data are part of the digital foundation in order to use ML and AI at the speed of mission, she continued. A strong data governance foundation and an enterprise perspective of data are also critical. Under the Federal Data Strategy Action Plan, the DAF published its Data Maturity Assessment Model, which has also been embraced at the DoD level. Another priority for the CDO is building data acumen at all levels, beginning with the creation of a pipeline of talent into the DAF. For example, the Air Force Academy now offers data science majors and minors to cadets, and the CDO partnered with the Air Force Institute of Technology to pilot a data science graduate certificate program. Other approaches to empower airmen and guardians at all levels include collaborative activities such as datathons. She also described a small data laboratory based out of Andrews Air Force Base, where any airman or guardian can enter a problem set and, if it has enterprise relevance, run a use case. The final piece of the DoD data strategy relates to interoperable and secure data, through the development of trusted partnerships and the eventual convergence of data platforms. She stressed that interoperability is the cornerstone of all CDO efforts and added that “vendor lock” has to be reduced or prevented for this strategy to succeed.

Because the CDO is part of the 2023 Program Objective Memorandum, it can make deliberate decisions about investments in data at an enterprise level moving forward. Ms. Vidrine detailed the progress that has been made on the VAULT data platform, the goal for which is to provide self-service analytics. Within only 100 days of conception, the data platform reached Impact Level 4; with agile development, additional capabilities are released every few weeks. One initiative of the data platform is “bring your own tool”; in other words, to meet a mission need in the field, it is possible to fast-track an emerging technology into the platform safely and securely through a partnership with the acquisition community and the security team.

Today, VAULT has reached Impact Level 6 with a cross-domain solution using “Infrastructure-as-Code,” which makes it possible to onboard airmen within hours.⁴

In closing, Ms. Vidrine described the DAF data fabric, the common architecture for department-level data. VAULT and other capabilities will be deployed on this data fabric, offering a truly interoperable solution. It maximizes return on investment, breaks down siloes, and automates data pipelines quickly to support the mission, with the added goal of reducing the number of standalone data feeds and creating automated data feeds for critical data sets. She explained that to achieve this vision, the CDO first has to identify which data sets are used most often in the DAF and which data sets are needed at the DoD level.

Open Discussion

Dr. Rama Chellappa, Bloomberg Distinguished Professor of electrical and computer engineering and biomedical engineering, Johns Hopkins University, asked how annotations that will be useful for downstream tasks are created as well as how experts are trained to work with an abundance of data. Ms. Vidrine replied that the CDO has a team of data scientists, data architects, and data engineers who work alongside functional experts to generate solutions, and she emphasized the value of having access to the right talent at the moment the mission need arises.

Lt. Gen. Hamel commended the CDO’s progress and inquired about remaining challenges and aspirations for the future. Ms. Vidrine accentuated that the technology component of the transformation is less difficult than the change management aspect. More “change champions” and “chief data evangelists” who invest their resources and staff to work on difficult problems are needed to accelerate change. She expressed her hope that people will stop thinking about how the Air Force operates today and start thinking about how the Air Force should be operating tomorrow by investing in the future. It is crucial that investments in legacy systems are eliminated, she continued, because such investments are too costly and do not accelerate change. She added that if the Air Force wants to retain its newest airmen, who are digital natives, it should embrace modern capabilities. Ms. Knausenberger explained that the Air Force has to select one of the following three paths: (1) change the organizational structure by aligning authority, funding, and competency and empowering a leader who is held accountable; (2) continue with the current coalition; or (3) fund a path forward. Although the Major Commands (MAJCOMs) recognize the value of having access to data and are eager to invest, the resources are not available (e.g., this year, the Air Force faced a $30 billion top-line reduction). Mandating enterprise solutions has not proven effective in the past, primarily because people do not realize how much up-front funding is needed to ensure that an enterprise service is ready for the fight. She surmised that most agree with the vision to transform and are ready to eliminate stovepipes, but an investment to enable this progress has not been committed. She emphasized that if a substantial investment to kick-start innovation is made now, a foundation could emerge, and billions of dollars would then be saved each year as expensive stove-piped secret networks that are open to many security threats are gradually eliminated. The likely most effective approach for transformation, she continued, is to target the Air Force’s organizational structure and authorities and improve the timing associated with funding. In response to a question from Lt. Gen. Hamel about whether money is being spent appropriately, Ms. Knausenberger said that money is not being spent on the right initiatives, not because people are being irresponsible but because people do not have the foundational enterprise systems to support

___________________

⁴ See https://cloudify.co/blog/infrastructure-as-code-is-it-really-enough-for-devops/:

An Infrastructure-as-Code (IaC) architecture is the management of networks, virtual machines, load balancers, and connection topology in a descriptive model. It uses the same versioning as a DevOps team uses for a source code. Like a key that only opens a single door, an IaC model generates the same environment every time it is applied. It is essential to the DevOps process and is used simultaneously with continuous delivery.

An example of how enterprises can use IaC is for software development. Programmers use IaC to create and launch sandbox applications. At the same time, QA professionals can run tests with perfect copies of product environments to test for errors. And then, when it’s time for deployment, you can push both infrastructure and code to production in one swift stroke.

current needs. She added that the Army and Navy are outspending the Air Force. Lt. Gen. Hamel wondered how much time it would take to redirect funds or create new investment streams to affect the desired transition, and Ms. Knausenberger remarked that real progress could be made in 1 to 2 years, as the chief information office has already identified billions of dollars in savings that could be realized by shutting down a few entities.

LEARNING FOR TRANSFORMATION: A PANEL PRESENTATION

Brig. Gen. Shawn Campbell, Deputy Human Capital Officer, Chief of Space Operations, noted that people have varied interpretations of what it means to be the “first truly digital service.” For guardians, a digital service incorporates the use of available digital capabilities and harnesses the power of data using digital platforms.

He described the development of the human capital strategy for guardians, which extends beyond simply generating and engaging talent. The first objective of the strategy is to connect in a collaborative, digitized environment, using a single platform for communication that is also connected to data sources for decision making. Guardians would be able to access this platform from any device while maintaining an appropriate level of security. The second objective of the strategy is to lead “digital enablement” by establishing a “digital cadre.” For example, when the Space Force has its first group of Supra Coders this year, a new set of skills will be available. The Space Force also plans to optimize its data infrastructure, because data will win or lose the next war. Brig. Gen. Campbell explained that it is crucial to be able to conduct data assessment in the human capital space—data on each guardian could be used to make decisions about gap analysis and to better align guardians and their skillsets with particular mission needs. The Space Force also aims to “supercharge” the way it automates processes, in particular fielding a talent operations platform that would digitize the talent management life cycle from recruitment through retirement.

A participant asked Brig. Gen. Campbell about undergraduate education for officer candidates. He replied that it is important for equitable talent to be distributed across both the Air Force and the Space Force. The U.S. Air Force Academy (USAFA) ensures that all cadets are exposed to space-related curriculum, which will help cadets to better determine whether they would like to pursue a Space Force specialty code. The Reserve Officers’ Training Corps (ROTC) is another area in which space programs could be incorporated. In Summer 2022, the Azimuth Program will expose cadets to space-related topics, and Officer Training School (OTS) recruits will be given an iPad prior to arrival to begin working on space mission sets. Gen. Holmes commended Brig. Gen. Campbell for this effort, emphasizing that the officer talent pool should not be limited to those who have chosen a career path at age 17.

Lt. Gen. Bill Liquori, Chief Strategy and Resourcing Officer, Space Force, noted that capability development processes are also shifting into the digital realm. Currently, DoD relies on capability development documents (CDDs) to generate requirements. These documents are ~100 pages and contain much repetition, owing to the standard template. As CDDs are reviewed at various levels, additional documents are created with commentary, which means that a final requirements document may not be generated for another 2 years. By this time, the threat has likely evolved and a different need has emerged. He advocated for a more streamlined process for capability development.

The requirements team and the Chief Technology Innovation Office team have begun to address this problem with the vision to create a “digital backbone” that would host a series of tools to generate “digital requirements packages.” In other words, Lt. Gen. Liquori continued, individuals would have access to multiple tools to build “artifacts” (i.e., sections of a digital requirements package). For example, a mission-area requirements package would have annexes for individual systems instead of requirements packages for each individual system.

Lt. Gen. Liquori explained that, ultimately, the goal is for the model-based systems engineering environment to move to this digital backbone so that during force design, the needed architectures and key performance parameters become apparent. This information could then be passed to the requirements team, creating an opportunity to eliminate much of the repetitive content in CDDs. People could coordinate

digitally in real time as this digital requirements package is being generated. Furthermore, with Space Systems Command on this digital backbone, insight about the future could be evaluated much sooner. He added that when Systems Command procures a system from industry, a digital twin will accompany it. This digital twin could replace a model with an actual digital twin, creating a time-saving cycle—force design, to requirements, to acquisition, and back. Another objective is for the Joint Requirements Oversight Council to have access to this digital backbone so as to provide ongoing oversight and maintain focus on strategic-level warfighting needs.

Lt. Gen. Hamel asked Lt. Gen. Liquori how to make better use of commercial and allied capabilities in force structures. He also wondered about the roles within the requirements process of those who will be expected to develop, acquire, field, and support capabilities. Lt. Gen. Liquori noted that international and commercial contributions occur during both force design and acquisition. In the force design stage, purely commercial, purely government-developed, and hybrid solutions are analyzed before a recommendation is made. With the creation of a digital backbone, the requirements team would have visibility into this ongoing work. Mr. Munson pointed out that (1) a commercial solution that matches some of the requirements will not be an exact match, and (2) commercial components from another set of approximate requirements cannot be modified without destroying the underlying economic model that supports the use of commercial products. He suggested adding a “go-back” step in the requirements process: once the commercial components become part of the reality, it is important to consider modifying the architecture and requirements, instead of leaving this problem for acquisition to address. Lt. Gen. Liquori said that there is no plan to mandate the use of commercial products for particular functions in a digital requirements packages. He noted that evaluation of commercial product availability occurs during the force design phase; in a situation in which a commercial product is identified as the best option, he would consider making adjustments to the requirements.

Ms. Gwendolyn DeFilippi, Assistant Deputy Chief of Staff, Manpower, Personnel and Services (Headquarters Air Force [HAF] A1), remarked that her team has been working to reduce the Air Force’s 118 disparate personnel management platforms to 5 to 6 interoperable platforms. In some cases, the adoption of commercial-off-the-shelf products is most effective; in other cases, customization is required. To address this challenge, her team partnered with an agile company to design a platform that supports the Air Force’s desired business changes. The primary objective of this new platform is to ensure that data interact and are accessible, which will reduce the instances of redundant or erroneous data input. She explained that plans for an integrated pay and personnel system are also under consideration. She emphasized the value of change management; it is the Air Force’s responsibility to communicate its needs to IT developers instead of expecting them to anticipate those needs and potentially create a process that will not work. She underscored the importance of both working with the community to think about ways to use technology to adapt an agile mindset and committing to iterate and improve incrementally (instead of setting full requirements).

Ms. DeFilippi described a significant barrier to developing digital talent: there is no specific functional authority that helps define requirements. Her approach to the digital talent development strategy is to have (1) executives with enough digital awareness to create environments that will facilitate a digital workforce, (2) mid-level staff with enough digital knowledge to navigate the digital space, and (3) a cadre of digital experts. She asserted that the Air Force will continue to struggle with a coherent digital strategy until it formally designates a lead for digital and cyber in HAF A1. Airmen will also continue to struggle until legacy IT platforms are retired—airmen need more modern technology that they can navigate.

Brig. Gen. Melissa Cunningham, Director, Cyberspace Operations, Deputy Chief of Staff, Intelligence, Surveillance, Reconnaissance, and Cyber Effects Operations, HAF A2/6, noted that in addition to the digital skills that are valued in the Air Force, storytelling and an understanding of programming and budget are key attributes. She reiterated that the new airmen arriving to the Air Force are digital natives, and the talent pool is exceptional. However, because the competition for this talent is significant, the services have to think about rebranding in a way that motivates people to join the organization. Another challenge is managing the talent that is already in the pipeline, which is currently done assignment-by-assignment. Retention is an issue for the civilian workforce as well. She emphasized that it is crucial to create a culture

in which civilians know that they are valued members of the workforce—in other words, train them so well that everyone wants to hire them but simultaneously treat them so well that they have no interest in leaving the organization.

Lt. Gen. Hamel asked about both the motivations and the frustrations that respectively attract and deter mid-level Air Force officers. Brig. Gen. Cunningham said that the most significant frustration for mid-level officers is identifying a career development path. Another challenge is when the spouses of military members have a career path with significantly higher earnings. It is critical to promote stability and convince family members of the value of the officer’s next assignment, she continued, because the most significant motivator for mid-level officers is the mission and the ability to defend the nation and maintain the trust of the American people.

EXECUTING TRANSFORMATION: A PANEL PRESENTATION

Lt. Gen. Tim Haugh, Commander, 16th Air Force (16 AF), explained that 16 AF operates and defends networks and supports Cyber Command to present offensive capabilities and operations in the information environment. He mentioned several areas in which his unit is succeeding. From an operational perspective, unity of effort is increasing inside of the enterprise based on the capabilities ACC has provided to work with data. A substantial investment has been made in a data platform that enables data aggregation across the network for more coherent defense. This creates consistency among the airmen that are defending the edge of the network, the airmen operating in the cyber protection teams, the mission defense teams, and the cybersecurity service providers at the weapons system level. Furthermore, 16 AF has created a network operations center and a security operations center in the 688 Cyberspace Wing—the Non-Classified Internet Protocol Router Network (NIPRNet), the SIPRNet, the Joint Worldwide Intelligence Communications System, and some Special Access Programs are aggregated inside of the 688 Cyberspace Wing, which also leads to more cohesive defense and operation of networks. The Air Force is also experiencing success with its cloud strategy: Cloud One and Platform One have increased attack surface and the ability to leverage data. Now that this foundation exists, he suggested that it begin to extend up networks and levels of classification.

Lt. Gen. Haugh also discussed areas in need of improvement from an operational perspective, including developing a better understanding of the threat and how it applies to the technical debt that the Air Force has created. Many of the Air Force’s weapons systems were built in a different world with a different threat—the adversary can now collect data on the United States every day as well as leverage commercial data to target the defense industrial base. He emphasized the need to think like a digital Air Force, creating and defending capabilities without allowing the adversary to gain advantage. To better confront the threat, he proposed securing weapons systems and ensuring that compute power is available at the edge to be resilient. He advocated for the development of key performance parameters for every new weapons system so as to avoid acquiring more technical debt and risk, both in cybersecurity and cryptography.

Gen. Holmes asked how mission defense teams operate at the wing level and coordinate with 16 AF. Lt. Gen. Haugh explained that the mission defense teams emerged as a strategy to buy down technical risk by dedicating a cyber-defensive team to the operational wing. Although the intent was to replicate this vision throughout the Air Force, resources are too constrained, and the initiative is now very targeted. He noted that there may be opportunities to use this construct for weapons systems, bringing the data back to more central locations where having a mission defense team is not feasible.

Col. Heather Blackwell, Director of Cyberspace and Info Dominance, ACC/A6, discussed the importance of instituting a data standard—for all data, not just cyber data—so that data can be collected on the data platform that Lt. Gen. Haugh described and used to inform decision making. She also noted the value of the Air Force’s cloud strategy and championed efforts to create a template to enable smart and secure migration to the cloud. Understanding risk is critical: decisions and initiatives need to be informed by existing cyber threats in order for the United States to remain competitive against the adversary. With unity of effort behind these ideas, she continued, the Air Force could become truly digital.

Col. Blackwell expressed her support for mission defense teams as a strategy for weapons system security, and she continues to push that effort forward. This approach helps to identify vulnerabilities not only in the weapons system but also in the components of the weapons system. She cautioned that adversaries can disrupt capabilities in air power in other ways beyond Internet protocol attacks. In closing, she underscored that culture change—investing in and educating people—is a significant part of digital transformation. As the push for digital transformation continues, it is crucial to focus on how to become more competitive against peer adversaries instead of focusing only on how to save costs.

Mr. Steven Wert, PEO Digital, Air Force Life Cycle Management Center, explained that the Air Force is making significant changes in how it acquires software—for example, with the emergence of software factories. Reflecting on the Air Force’s current acquisition processes, he noted that because industry partners use applications and models that the government typically does not own, paper reports are produced. In order for the government to conduct analysis, this information then has to be entered into models manually. Industry’s networks are not connected to the government’s networks, and the government uses outdated tools (e.g., Microsoft PowerPoint, Word, and Excel spreadsheets) instead of relying on automation for assistance. Mr. Wert stressed that using applications that do not communicate and airgapping networks as a security strategy (rather than using identification- and authentication-based access) will not equip the force to win a 21st century conflict.

Mr. Wert compared the success of agile DevOps for software to the potential of digital transformation for more hardware-intensive systems, all of which are operated with software. He explained that digital execution for hardware begins with having access to data, source code, and tools. The Air Force’s working relationship with its industry partners would need to change fundamentally to enable this transformation. He noted that agile DevOps changed the Air Force’s approach to requirements—a recognition of the important “requirements” among users and operators initiated new methods for contracting, testing, and funding; digital acquisition will have similar impact and will require a whole-of-Air Force transformation.

Currently, requirements development and programming begin several years prior to funding; requirements are fully defined before contracting and must be fully met before a weapons system is delivered for operational acceptance. This method delays the fielding of capabilities, and immediately after the system is delivered, it is modified and the process continues. He explained that digital acquisition, which requires a different way of thinking about programs, could radically change this process. He advocated for running thousands of iterations using digital models to discover what requirements could or should be instead of fully defining requirements up front. This would generate a better understanding of the cost implications of design traits as well as allow end users and government engineers to have a voice in the design process. He emphasized that the Air Force is “behind the curve” because its funding and execution process is too slow, and he reiterated the value of having a continuous focus on the next design rather than on the production and sustainment of the current design.

Mr. Wert mentioned that his directorate has been successful with modern software practices: it relies on the defense industrial base more as a service, has changed its relationship with industry, and has experience implementing open architectures. Remaining challenges for the Air Force more broadly include addressing workforce issues, maintaining business relationships, and embracing change at scale. There is also a need to expand the size and skillsets of the government’s technical workforce to enable hands-on modeling, simulation, and analysis, he continued. An ongoing related effort focuses on strengthening partnerships with industry and federally funded research and development centers, as well as collaborating with university affiliated research centers. He emphasized that the magnitude of the change required for digital transformation is perhaps larger than what was required for agile DevOps, because it demands modifications to requirements, budgeting, testing, contracting, and logistic support processes. In closing, he stressed that acquisition strategy approval requires understanding how agile software, open systems, and digital engineering will be implemented.

Open Discussion

Mr. Santee resonated with Mr. Wert’s commentary on the urgent need for agile solutions to everchanging requirements with integration into an open architecture. In order for the Air Force to achieve a different outcome, he continued, it is essential to change the current approach to requirements, resourcing (i.e., budgeting and execution), and acquisition and to create a data-driven environment. He asked Mr. Wert to share lessons learned that may apply to the DAF’s transformation journey. Mr. Wert reiterated that agile DevOps resulted in large-scale changes—digital engineering can have a similar effect. He emphasized that each PEO should have the authority to drive change, as was the case for agile software development.

Dr. Ryan explained the concept of the digital twin, in which data from the operational twin feed back into the digital model to increase understanding of how operational parameters are affecting the performance of the operational twin. Once problems are addressed in the digital model, it is possible to roll out upgrades. However, the notion that an adversary could modify the data so that the data being transferred over the thread to the digital twin are not representational of the operational twin is very dangerous; she cautioned the Air Force to be mindful of how this could ultimately cause the operational twin to become vulnerable to attack or to fail.

Mr. Munson wondered how the relationship with the defense industrial base is evolving. Mr. Wert responded that although the progress has been slow, it has been beneficial to have a service acquisition authority to challenge industry on its business models and intellectual property strategies. Strong leadership will also be key in the Air Force’s transition to digital engineering.

Lt. Gen. Hamel asked Lt. Gen. Haugh whether he has command authority over mission teams and if he can advocate for capabilities. Lt. Gen. Haugh replied that in its defense of the Air Force’s networks, 16 AF has authority to direct a change in the configuration of a network, add defensive layers, or remove someone from a network, based off of the threat or guidance from the Joint Force Headquarters DoDIN Commander. Changes to mitigate network threats are executed through Operation Cartwheel: a unity of effort among airmen that aligns defensive activities—starting at the network edge down to the individual elements at a base level—that are configuring the network to operate and then layers defenses with the mission defense teams and cyber protection teams. In addition to delivering networks and network defense for the Air Force and the Space Force, 16 AF also conducts offensive and defensive cyber space operations and operations in the information environment in support of multiple Combatant Commands. Lt. Gen. Hamel wondered if provisioning is still the responsibility of the respective MAJCOMs and the Space Force, in terms of their mission systems or their base-level capabilities. Lt. Gen. Haugh responded that, from the overall network perspective, ACC is the lead for provisioning; 16 AF is involved in the operational readiness reviews before the capabilities reach the network to ensure that they are executed in accordance with the guidance of the chief information office and Cyber Command. Although gaps remain in configuration control, as well as in roles and responsibilities, 16 AF is working tirelessly to ensure resiliency. Col. Blackwell added that ACC works closely with the operators to understand their priorities for cyber weapons systems and then advocates for funding lines. ACC also ensures that the MAJCOMs understand the overall constructs of the big data platform and of command and control.

In response to a question from a participant about Air Force culture, Col. Blackwell said that culture changes begin at the Cyber Schoolhouse, where officers and enlisted are taught not only about devices but also about how to be operational problem solvers who can conduct briefings and debriefings and embrace critical feedback. The Cyber Schoolhouse now offers modular lessons so that recruits, who already have digital talent, are learning at the speed of cyber. Testing is another area that would benefit from culture change; Col. Blackwell advocated for using the test environment and sandboxes to test capabilities in a rigorous model before integrating them on the operational network. She asserted that agile solutions and a better articulation of risk in funding and threat vectors are also key to demonstrating the consequences of not investing.

Mr. Munson inquired about the contrast between a democratized service of digital natives with access to data and connectivity, and the traditional chain of command structure of the Air Force. Col. Blackwell explained that even though the data are available, service members are still expected to comply with the

orders they are given and work within their purview. She expressed confidence in this new realm of data transparency, owing to clear commander’s intent and accountability for those who abuse authority or data access. Lt. Gen. Haugh added that he has not experienced any issues with airmen engaging inappropriately with the data. Mr. Wert noted that lieutenants and captains are now able to implement and execute change, in part owing to their increased coding abilities. Given the value that these individuals have to companies beyond the Air Force, he emphasized the importance of considering retention strategies for this talent—if they do not see a career path that includes innovative work, they will leave the Air Force.

Ms. Westphal asked about the possible role of chaos engineering⁵ in the digital transformation. Dr. Nielsen responded that Netflix created a tool a decade ago, Chaos Monkey, to ensure that its system was resilient: by randomly turning off parts of the system’s software, the tool tested how well the system could work around issues in its applications and services. This practice of looking for both vulnerabilities and resiliency by testing for random failures in major systems is one that the DAF should consider, he continued. Such tools are constantly evolving to keep pace with the changing environment, and they are either free or very inexpensive to use under a license (i.e., the Air Force could lease them without needing to maintain them). Mr. Wert explained that in its experiments with chaos engineering, Kessel Run⁶ discovered a sense of resilience and revealed a way to do maintenance without taking down a system. Dr. Nielsen added that Amazon, Netflix, and Google are constantly (and successfully) implementing system changes while operating.

Gen. Holmes asked Dr. Nielsen to talk about digital modernization in the context of software factories. Nielsen replied that the modern way to build software is via a “software factory,” which creates an environment of productive programmers by automating routine tasks such as configuration control, version control, and nightly security testing. This technology approach was developed in the commercial world, and the military-based software factories currently in existence were initially conducted as experiments, with Kessel Run as the first. DoD considered how to scale up from these experiments and learned that it is important to let each software factory specialize in particular systems; standardization has the potential to freeze technology in place, which is undesirable given how quickly software engineering technology changes.

Gen. Holmes pointed out that it is difficult to find money for updating the Air Force’s legacy systems within a construct built to assign money to programs that primarily focus on hardware. He wondered how to balance resources so as to have a dedicated, predictable stream to recruit and retain talent without having those people waiting between projects and searching for jobs with other organizations. He said that more incentives are needed to motivate people to have careers in the Air Force and share their expertise to help DoD compete with China. Dr. Nielsen noted that it is motivational for people to see their work making a real-world difference (in contrast to the more traditional acquisition system, in which even after years of work, a product might never be developed).

WORKSHOP ONE CLOSING REMARKS

Lt. Gen. Hamel invited planning committee members and workshop participants to share their observations and takeaways from the first workshop of the series. Gen. Holmes observed that the infrastructure required to support digital modernization, as well as the related bandwidth and technologies, had not been discussed. Lt. Gen. Hamel added that balancing the maintenance of legacy systems with investment in the future is a long-standing problem for the Air Force. Gen. Holmes noted that people often choose only to blame Congress for its denial of funding requests; however, the Air Force has to overcome its tendency to buy more things and focus on making things work. Lt. Gen. Hamel pointed out that industry has similar issues, but Gen. Holmes emphasized that industry charges its customers. Recalling Mr.

___________________

⁵ Chaos engineering is the discipline of testing the resiliency of a system.

⁶ U.S. Air Force, 2021, “Kessel Run Delivers Chaos Engineering Practices to Black Pearl,” Press Release, https://kesselrun.af.mil/news/CHAOS-Engineering.html.

Reardon’s and Mr. Lombardi’s presentations, Gen. Holmes reiterated that the Air Force is not currently organized to direct and implement a major transformation: doctrinal and organizational changes are needed to successfully leverage the abundant data available and the tools to connect them.

Ms. Westphal shared key takeaways from the first two days of the workshop series—most importantly, the Air Force has a lot of collaboration without a unity of effort. She observed that there is a high usage of data without representation of how those data could be used; the Air Force cannot increase speed or efficiency if data are being layered onto an old structure of operations. It is also important to identify and analyze risk, both today and for the future. She advocated for chaos engineering as one step toward modernization, because achieving resiliency in an everchanging world is key. She cautioned against creating large programs with significant funding that do not achieve intended outcomes, and she recognized that people (i.e., developing a culture of practice and exercise) are an important component of the DAF’s digital transformation.

Lt. Gen. Hamel remarked that the existing rigid ecosystem and processes of the Air Force are misaligned with the nature of digital technology and its evolution; resource allocation, skillsets, and acquisition all have to be approached differently. Dr. Nielsen added that the requirements engineering process perfected over the past 30 years has hurt the Air Force: it takes 5 to 6 years for a capability to emerge, and by that time the technology and the threat have both changed, rendering the capability outdated. He said that it is important to design more adaptable systems that evolve, just as commercial software companies do. Gen. Holmes agreed with Dr. Nielsen that the Air Force has an opportunity as it begins to acquire new things in new ways to change the process; instead of trying to fit software systems into the old model, it is time to create a requirements process that both works for software-intensive systems and casts back to other systems. Dr. Langston noted that the tools for software-intensive acquisition are already available, but the budgeting process cannot be fixed without congressional assistance. He suggested building budgets that would allow the chief information office and the CDO to leverage funds as seed money to add to other people’s resources, which would create a unity of effort across the enterprise. Dr. Green said that “competent” (instead of “capable”) people should work on requirements. Ms. Westphal added that definitions of words like “requirements” vary across the enterprise, which makes it difficult to achieve change. Gen. Holmes pointed out that although the requirements process currently is driven by congressional oversight and fairness in the contracting system, it should be driven by the need to compete with China.

Dr. Chellappa commented that “digital transformation” is also a general term with several interpretations. He added that if the DAF continues to forgo relationships with the nation’s best companies and top talent, the United States will be at a disadvantage to China. Mr. Munson noted that China operates differently than the United States because it has a state-run capitalist system. The United States does not have industrial policy or industrial planning; U.S. industry operates primarily on one of two business models, and neither is well matched to the way the government operates. Although the nation’s best companies could help the DAF, their business incentives keep them from doing so. He explained that the Air Force has yet to determine how to create an entirely new structure and culture to support data, despite the common belief that data will win the next war, and he reiterated that the notion of a horizontal data architecture will challenge the traditional notion of commander’s control. Col. Douglas DeMaio, 187th Fighter Wing Commander, Alabama Air National Guard, pointed out that the same concerns surfaced when the concept of multi-domain emerged, and doctrine clarified how airmen would act as well as how to move forward with mission orders alongside the vision to accelerate change or lose. He suggested that the same approach be taken with digitization—define and issue direction in doctrine, and balance top-down and bottom-up approaches that set the pace for the Air Force.

Lt. Gen. Ted Bowlds (USAF, ret.), chief executive officer, IAI North America, expressed concern about the Air Force’s disparate activities and lack of strategy and governance. Current contracting requirements hinder the Air Force’s ability to work with talented individuals from industry. He cautioned that the tools to manage data (e.g., AI) do not seem to be moving forward in parallel to the increasing generation of data, even though these tools are needed to help focus the data for the decision maker. He emphasized the need for the Air Force to plan for continued operations in a disconnected environment.

Mr. Santee observed a philosophical culture clash between the younger and the more seasoned members of the Air Force in terms of their commitment to digitization and transparency. Without a unity of effort, he continued, it will not be possible to achieve a digital transformation. For example, if the Air Force believes that information is going to win the next war, why would it purchase another platform? Instead, he suggested that the Air Force find, pool, share, and link data in the cloud, and reallocate limited resources. He advocated for embracing the digital talents and contributions of the younger generation and empowering them to make decisions. He raised a question about the relationship between radical data transparency and security and also wondered if this new level of transparency might change the relationship with Congress for the better with increased information access. He added that force structure will also experience changes as a result of the digital transformation; for example, the Space Force plans to have far fewer people than the other services but still accomplish the same tasks by utilizing data and applications for automation.

Ms. Westphal referenced a 2011 Center for a New American Security lecture about why generals in WWII were more successful than those in Korea, Vietnam, Iraq, and Afghanistan. In WWII, if generals did not achieve within 90 days, they were relieved and moved to a different task where they could be more successful. Now, leaders remain in positions even when they are not successful, and so the motivation to achieve and transform is lacking. Gen. Holmes clarified that most WWII generals were warfighters. He noted that the centrally directed culture of today’s Air Force is problematic because processes are driven by accountability and oversight instead of by reward for producing results beyond the norm, stifling individual initiative. Ms. Westphal pointed out that the culture of decision making has to change; new architectures and technologies alone will not solve this problem. Lt. Gen. Hamel added that leadership and collaboration are important components of the success of digital transformation. He described the need for a common conviction about what is essential for the future of the Air Force.

Mr. Drolet acknowledged Mr. Lombardi’s concern about the variation in the definitions of “digital” but added that sometimes progress is stalled by an overemphasis on the need to define concepts. The Air Force has to be agile, but technology is not the barrier: the current bureaucracy, culture, policies, and processes tend to slow progress. He noted that the Air Force has much farther to go to achieve unity of effort, and he suggested an increased focus on the importance of its people in the digital transformation.

Lt. Gen. Hamel reiterated that despite several innovative efforts, the Air Force remains uncertain about how best to attack the problem of its organizing principles. It is difficult for an organization as large as the Air Force to think from an enterprise perspective, especially without the dedication of more resources, commitment, and talent to this challenge. He emphasized that organizations across the world confront similar problems as the digital revolution continues.

Dr. Green noticed a limited discussion of digital representation during the first workshop. She stressed the importance of understanding that democratization means having a common language and common understanding. It is also necessary to consider how to separate digital talent from processes and procedures, and determine what tasks can be automated and what tasks can be augmented to best leverage intellectual capital. She explained that bright young people often leave the Air Force because they do not know where they fit in—they are not shown the “big picture”—and it is crucial to create pathways for the development of new skills and how they will align with the organization.