The Joint Training Enterprise (JTE) requires effective integration of and technical interoperability among disparate synthetic training capabilities from across the Services to enhance joint operational capability and achieve joint readiness. Opportunities to enhance joint training interoperability increase when disparate synthetic training capabilities employ common or shared models and simulation data.
This idea underpins the development of common data services (CDS), a coordinated capability designed to rapidly locate, access, transform, transmit, to enhance the JTE’s synthetic training capabilities. Overlaying the CDS concept across the JTE creates the framework for Joint Federated Common Data Services (JFCDS), which enables each Service to develop and maintain its own data service provisioning capability, federated through common technical standards and protocols, which together allow the sharing of authoritative source data among the Services and across the JTE.
This article explores data-related synthetic training interoperability gaps, considers how current capabilities and capabilities in development (partially) address these gaps, and shows how JFCDS effectively leverages the successful attributes of these programs while meeting remaining data-related shortfalls.
While this paper focuses on the needs and solution for joint training, the same needs exist for training within the Services and extend to other application areas for modeling and simulation with the DoD. The modeling and simulation users supporting acquisition, experimentation, and test and evaluation require the same type of CDS solutions to provide current, authoritative, and appropriate data. The ability to locate, obtain, and integrate data from sources across DoD components is critical to enable innovation at the pace necessary to provide agile and adaptive systems for the defense of our nation.
The Joint Training Enterprise requires a systemic approach to synthetic training capability development and integration that leverages collaboration and cooperation among joint training stakeholders by enhancing, facilitating, and synchronizing information sharing, capability development requirements management, and technical interoperability. A series of documents articulates DoD and Joint strategy, policy, and governance to this end, including DoDD 5000.59, “DOD Modeling and Simulation Management”, DoDI 5000.70 “Management of DoD Modeling and Simulation (M&S) Activities,” the 2018 Joint Technical Training Interoperability Strategy, and the 2016 Joint Training Technical Interoperability Memorandum For Record. Yet for this approach to succeed, Joint stakeholders must commit to the development and Joint-level integration of their synthetic training capabilities so the Joint Force can truly train as it fights. When disparate synthetic training capabilities employ common or shared models and simulation data, the capacity to integrate these capabilities increases. The Joint Training Enterprise, led by the Joint Staff J7 through its Joint Training Synthetic Environment (JTSE) Work Group (WG), has therefore an interest in developing a framework for the discovery and access to authoritative data sources.
This initiative seeks to provide common data services, a coordinated capability to rapidly locate, access, transform, transmit, and distribute authoritative source data. These data types include terrain and geospatial, order of battle, parametric, and operational environment data describing Political, Military, Economic, Social, Information, Infrastructure, Physical Environment, and Time (PMESII-PT) variables — each correlated into application-usable formats that facilitate interoperability among joint synthetic training capabilities. This article will describe the simulation data-related problems and gaps within the broad portfolio of synthetic joint training enablers and show how the concept of common data services—implemented through the provision of a standardized architecture of authoritative data and transformation as a service in a cloud based, web enabled platform—will facilitate interoperability among these capabilities across the Joint Training Enterprise. Accordingly, this paper proposes a framework for Joint Federated Common Data Services (JFCDS) in which each Service develops and maintains its own data service provisioning capability, federated through common technical standards and protocols, which allows the sharing of authoritative source data among the Services. JFCDS leverages individual Service expertise on their respective authoritative data sources and training enabler data needs for the benefit of the whole Joint Training Enterprise. JFCDS further improves synthetic training capability interoperability across Services and therefore enhances the training of joint combined arms operations.
Joint training exists to improve joint operational capability and to achieve joint readiness, each integral to the Services’ collective capacity to conduct joint combined arms operations. The conduct of joint training exercises, using distributed simulation-based training enablers, constitutes an integral component of the joint training strategy. These simulation-based training enablers depend upon readily accessible, consistent, up-to-date data in both their development and subsequent employment across the Joint Event Life Cycle (JELC)—this includes event planning, scenario generation, exercise execution, and after action review (which correlates to the Army’s Operations Process—Plan, Prepare, Execute, and Assess). Unfortunately, the JELC cycle is often costly in terms of time, manpower, and resources due to the difficulties involved in synchronizing and integrating the current portfolio of joint synthetic training capabilities in service of a common joint training event.
Yet the Services’ use of the same authoritative data baseline for their respective synthetic training capabilities (present and future) may significantly reduce joint training interoperability challenges. Additionally, such a data-service foundation would enable data preparation and management throughout the JELC, thereby reducing time, manpower, and resource requirements.
Problem Overview and Root Causes
With respect to the data required to support the use of Live, Virtual, Constructive, and Gaming (LVC-G) simulations and capabilities in conjunction with mission command/command and control systems (MC/C2), three core problems currently exist. First, the preparation and provisioning of appropriate data for use LVC-G simulation and MC/C2 systems for joint training (and Service-specific training) takes too much time. Using current methods and tools (and in response to often changing exercise design and associated training requirements), data preparation and provisioning can take weeks to months. Second, these data preparation and provisioning processes are labor-intensive (requiring a significant, dedicated staff) and thus are costly. Finally, inconsistencies in data used among LVC-G systems and between LVC-G and MC/C2 systems negatively impact training quality, consistency, and availability.
The above core problems are attributable to a variety of root causes. In some cases, governance bodies have not identified the appropriate sources for data required for joint training. Data sources include both the authoritative sources and the sources for additional data required to provide sufficient level-of-detail (LOD) for applying LVC-G simulation. Additionally, authoritative data sources are often hard to access (security or policy issues), incomplete in content; or include errors, gaps, and out-of-date data. Training and exercise planners cannot consistently and thoroughly discover what data is available from these sources, as well as what data from prior training events is suitable for reuse.
Once data has been identified, automated capabilities are lacking at times to support request and delivery or direct retrieval of the needed information. Complicating the use of data is the fact that there are many formats and standards for simulation data – currently each model and/or simulation has proprietary data formats and standards. Simulation data managers often create and maintain characteristics such as parametric, probability of hit/probability of kill (Ph/Pk), weapons pairing, etc., at the individual simulation level. Some of this variation reflects the current reality that Services have different simulation needs and training requirements that demand varying levels of fidelity and resolution.
Training and exercise planners lack automated capabilities to combine and transform data from multiple sources into a form that is appropriate for use in LVC-G and MC/C2 systems. These features include capabilities to select data subsets, merge data from multiple sources, and transform data both semantically (e.g., enumeration translation) and syntactically (e.g., format translation).
As seen, data challenges encompass issues with “visibility, access, extraction, understandability, trust, interoperability, transformation (including [modification], fusion, integration, enhancement, filtering, and tailoring), and reuse” (PEO-STRI, 2015). The aforementioned issues clarify the problem space and inform (practically and methodically) the requirement for both technical and management solutions that incorporate JFCDS.
Foundational Data Strategy
The data concerns of the US Government and DoD are far larger than how data supports Joint Training; nevertheless, the US Government’s strategy to solve difficult data-related problems will inform this paper’s recommended approach. Strategic documents that shape the broad approaches that the Federal Government and DoD employ in their respective enterprise data strategies include the DoD Net-Centric Data Strategy (US Department of Defense, 2003), DoD Net-Centric Services Strategy (US Department of Defense, 2007), and the Federal Cloud Computing Strategy (Kundra, Vivek: US Chief Information Officer, 2011). The Joint Federated Common Data Services framework proposed in this paper seeks to build on these strategies and nest with the goals of the JTSE in order to build a coherent, holistic approach for managing the provision and use of data among emerging synthetic training capabilities.
With these core problems, root causes, and governing strategy documents as a point of departure, the authors present the following vision for Joint Federated Common Data Services:
The Joint Training Enterprise will, in conjunction with Intergovernmental, Multinational, and Commercial partners, develop the technical and procedural infrastructure required to ensure the availability of data and enable its rapid discovery and retrieval; while leveraging common data service principles and standards; and employing Authoritative Data Sources (ADS) in standardized “simulation and application agnostic” exchange formats, to inform the development and use of next-generation, interoperable training capabilities within the Joint Training Synthetic Environment (JTSE), including the Army’s Synthetic Training Environment (STE), the Navy Continuous Training Environment (NCTE), the Marine Corps Synthetic Training Environment (MCSTE), the Air Force Operational Training Infrastructure (OTI), and other emerging Service synthetic training capabilities.
Current Capabilities and Capabilities in Development
This section considers the ways data managers in Joint Training and other M&S application areas have positioned past, current, and in-development data management capabilities to meet requirements, while exploring remaining data-related gaps. This analysis seeks to inform design discussions and Joint Title X authorities for validation of emerging data-related capability requirements that the JFCDS is intended to satisfy. Over the last 25 years, data managers in DoD that have sought to address many of the issues outlined in above, via numerous efforts. These efforts focused on one or more of the following activities:
- Discovery of data for use in M&S.
- Transformation of data from its source form and the merging of data to generate data appropriate for use in M&S.
- Access to (or retrieval) of that data.
These activities support three high-order, training simulation use-cases. First, data is need for exercise planning and preparation. Second, software engineers need data to support the engineering the synthetic training capabilities. To reuse of data and support event timelines, commonly a third preemptive activity occurs–the bulk preparation of data to support those first two activities.
These efforts inform both design discussions and Joint Title X authorities for validation of emerging data-related capability requirements that the JFCDS is intended to satisfy. A survey of both current and future capabilities demonstrate how each are positioned to meet data-related requirements while exploring remaining data-related gaps. Programs like the Global Force Management Data Initiative (GFM DI), Defense M&S Catalog, Joint Data Support (JDS), Joint Rapid Scenario Generation (JRSG), Joint Training Data Services (JTDS), Enterprise Data Services (EDS), Data Services Environment (DSE) and Unified Data have made important strides in the provision of a variety of disparate data sources. To varying degrees, they promoted reuse, provided analytical baselines of models and data, and showed that governance was possible through the establishment of standards for databases, scenarios, and terrain. Innovative features such as semantic search and scenario development made finding and using data easier. Additionally, AMSO’s Unified Data initiative, in its design to acquire access to authoritative source data, constitutes another important achievement. These programs also demonstrated limitations of keeping source data and model catalogs current, providing data at the required level of accuracy and resolution (i.e., data abstracted at too high of a level), and answering all data questions with only metadata. These limitations continue as manifestations of the core, systemic problems in data provisioning.
Here the authors describe the Common Data Services concept that we advocate advancing within the Joint Federated Common Data Services framework. The data required by today’s M&S systems spans a broad range of uses, type, content, and resolution. Managing this data, and providing timely and effective access to it, requires leveraging and advancing key capabilities and technology areas, including:
- Architecture (systems, reference)
- Standards (data exchange, enumerations, logical data models/ontologies, service interface specifications)
- User interfaces (user applications, widgets)
- Tools (transform, inspect, auto-correct, enhance, integrate/fuse, add value, tailor, auto data creation)
- Data tagging (discovery, structural, and semantic metadata)
- Discovery tools and services
- Automatic service composition and orchestration
A coherent architecture to support the data services’ operational activities—from data generation, through data integration, provisioning of data to simulations, and managing the data produced for and/or resulting from M&S executions—is essential. The architecture must rely on the use of standards to reduce integration costs by ensuring interoperability at the technical, semantic, and syntactic levels. The architecture must utilize a services-based approach to the maximum extent possible; we must therefore customize data management capabilities through integrating and orchestrating a set of commonly available services. Reuse or development of appropriate system-level architectures is essential for ensuring different M&S users (and different M&S systems) can access, exchange or retrieve, and understand the data they need. This requirement demands significant use and/or extension of appropriate standards that ensure consistency and cost-savings in handling diverse data. These include standards for accessing, identifying, representing, understanding, and—importantly—transforming data according to established semantic, syntactic, and technical standards that are agnostic to the ingesting simulation or application. Accordingly, data management should be a cooperative, Service effort that leverages open standard formats and machine interfaces, to provide the user required data access at the point of need (PON).
Identifying the relevant and desired data relies on the use of proper data tagging and discovery techniques. This process, in turn, requires the development of meaningful and standardized tagging techniques and encoding values that can be processed by machines (and not just humans), as well as the engineering and development of methods for automating both the tagging and discovery processes. The development of supporting tools and services that enable automated data tagging and discovery is also essential. Rapid and accurate retrieval of data, using standard discovery and structural metadata, constitutes an existing gap in need of addressing.
The development of common data services includes the technical and engineering efforts required for the deployment of the tools and services. These include integration with existing systems and services, understanding and accommodating the capabilities and limitations of existing data warehouses, repositories, and consuming systems. The essential elements in automating data handling processes are tools and services that can operate reliably and with minimum or no human intervention, extract the required information needed by specific consuming applications, present the data in the desired form. These tools and services may perform a variety of tasks including inspection, transformation, auto-correction, enhancement, integration / fusion, value adding, tailoring, auto-generation, identification, and discovering the required data. Common services also require well-designed interfaces and protocols to be intuitive to users. These interfaces and protocols must also be capable of detecting and self-forming to automatically connect the sequences of data operations to produce a chain of processes that meet specific requirements for handling, modifying, or transforming the data.
In light of the aforementioned root causes and core data-related problems; US Government, DoD, and Service data strategy guidance, and the vision for a potential solution that centers on a common data services concept, this paper advocates the development and implementation of the following:
That each Service and the Joint Staff (and potentially coalition partners) develop and maintain its own data service provisioning capability, federated through common technical standards and protocols, that enables the sharing among Service (and coalition) partners (through proper channels and at the appropriate classification level) of authoritative, standardized, data in exchange formats that are agnostic to simulation or application.
This approach reflects the reality that there is currently no validated requirement for common data services at the joint level (nor resources specifically aligned against this effort). Therefore, JFCDS will rely on the Services (with support from the Joint Staff) to establish a framework to provision and share their own data services amongst one another. This Service, common data service sharing will function through a mutually agreed upon set of standards, policy, protocols, and data exchange agreements. Critically, this article does not recommend the implementation of a particular CDS data exchange standard, format, or application programming interface (API) that would risk obsolescence. Instead, we propose an ongoing evaluation of open standards favored by commercial and industry leaders that are capable of evolving in a way that is commensurate with rapid technology advancement. This approach ensures that JFCDS remains robust and will not become obsolete even in the face of significant technological advances.
The Joint Federated Common Data Services (JFCDS) framework allows for several significant advantages:
- Flexibility. This approach allows each Service to develop its own data provisioning capability at its own pace as requirements and resources align. The authors stipulate that each Service will pursue next generation synthetic training capabilities at its own pace based on considerations like validated requirements, available resources, and training strategy. For example, the Army is aggressively addressing its data service requirements in the development of the Synthetic Training Environment’s (STE) Training Simulation Software (TSS) and Training Management Tool (TMT). Accordingly, Joint Federated Common Data Services can begin delivering capability in piecemeal, as individual Service data provisioning services come up on line (e.g., Army Common Data Services) and become federated to one another.
- Data Expertise. Services are experts on their own data requirements and source needs (e.g., force structure, parametric). Accordingly, each Service is uniquely positioned to identify and access the suites of data needed to support its respective training enablers (and therefore, which data sets to share among fellow Services).
- Leverage J7 Leadership. The Joint Staff J7 sponsored Joint Federated Common Data Services WG (as part of the broader Joint Training Synthetic Environment WG) can continually help inform common joint standards for data discovery, access, extraction, transformation, distribution, and transmission among Services. This enduring WG can inform and influence individual Services as each develops its own respective, capable data-provisioning service to ensure high levels of interoperability and data-sharing capacity among each Service.
JFCDS Application Contexts and Functional Overview
The Army provides an example of what a Service can do to implement common data services. The Army envisions its next generation Synthetic Training Environment (STE) as a cloud-based distributed system. The STE Cloud is defined as a set of application services, common data services, and supporting infrastructure. Army Application Services include the Training Management Tool (TMT) supporting planning, preparation, execution and assessment activities; and Training Simulation Software (TSS), which supports runtime activities. Army Common Data Services (ACDS) provides a foundation for application services to extract, transform, load, and distribute data from Army authoritative data sources and between STE Cloud instances. Supporting infrastructure is a key enabler for the STE Cloud.
Expanding the approach depicted in the figure above provides the groundwork for our proposal that each Service, Joint, and Coalition partners establish solutions for accessing authoritative data source and managing data needed for engineering of training systems and for conducting training with those systems. As each Service and partner establishes their respective common data service solutions, the Common Data Service Working Group should collaboratively identify the service-interface standards and specifications as well as Service Level Agreements (SLAs) needed to enable a meshed access to each other’s authoritative data sources and exercise data—insofar as that information needs to be shared. These interface standards and specifications are critical, as JFCDS must avoid the “n squared” problem of requiring new, additional, tailored gateways to enable data sharing and exchange among each Service’s respective synthetic training capability system architectures.
To achieve the necessary data infrastructure, a working group/integrated product team (IPT) (including service, joint, and eventually coalition stakeholders) must select and set service standards for the following:
- discovery (search and subscription)
- retrieval (request-response, publish-subscribe, or other means of delivery)
- access control (user, group, and attribute-level authorization; cross-domain solutions)
- configuration management and version control, and
- collaborative data management (e.g., in support of exercise planning)
Service and data exchange standards adopted by JFCDS will be consistent with the DoD Standardization Program, relying on open, consensus-based standards where available, and establishing military standards only when appropriate open standards are not yet available.
Services, joint, and coalition partners that employ JFCDS may share force structure, environmental, characteristics/performance, plans/operations, and other types of data in support of exercises, thereby assisting to establish sufficient interoperability among the respective training environments. Moreover, while not all data sources are needed by all stakeholders, common integration standards will ensure that data may be accessed anywhere as a need arises.
Conclusion and Way Ahead
This article documents the collective challenges that Combatant Commands, Services, and Agencies face in the effort to produce common data services that are available to the Joint Training Enterprise for reuse and interoperability–all to support joint readiness and the capability to execute effectively joint combined arms operations. Additionally, challenges remain as joint and Service combat M&S capability developers work to coordinate a synchronized joint training capability development strategy that promotes cooperative and collaborative development, prevents unnecessary redundancy and stove-piping, and identifies standardized technical and procedural approaches. The implementation of Joint Federated Common Data Services represents an important component of the broader joint training capability development strategy as JFCDS will—through the Services and with support from the Joint Staff—establish a framework for the Joint Staff and the Services to provision and share their own data services amongst one another.
Future efforts may include JCIDS-like efforts such as Capabilities Based Assessment (CBA), and defining measures of performance such as reuse, access, and interoperability that ultimately leads to CDS Standardization and JFCDS implementation. Other near term actions include the development of standardized, simulation agnostic data models for terrain, force structure, and entities for Services to employ in joint synthetic training; as well as the arranging of access and retrieval permissions for authoritative data sources from each Service and from the JS J7 that can populate the standard data models under development. Authoritative Data provisioned by Joint Federated Common Data Services supports the replication of the complex operational environment at a high level of detail across the air, sea, land, space, and cyberspace domains; replicating operational variables and mission variables (PMESSI-PT and METT-TC) and elements of national power (DIME).
While the authors focused on the needs and solution for Joint training in this paper, the commonality of needs, problems, and root causes, and solutions extends far beyond that Joint Training. These are common to training at large, as well as to other activities that employ modeling and simulation with the DoD. The needs and solutions from training most directly relate to those for acquisition, experimentation, and test and evaluation, but also extend to other applications such as strategic analysis. Common, consistent, authoritative data must drive all these activities to support innovation and agility.