Research Highlight: Enterprise Architecture Guide Development Support

The SERC portfolio of funded research tasks furthers its guiding vision of being the networked national resource to further systems research and its impact on issues of national and global significance.

Research Highlight

• Research Task: DoD Enterprise Architecture Guide Development Support
• Sponsor: Office of the Under Secretary of War for Research and Engineering (OUSW(R&E))
• Research Team
  • Principal Investigator (PI): Dinesh Verma (Stevens Institute of Technology)
  • Co-Principal Investigator: Tom McDermott (Stevens)
  • Key Contributor: Valerie Sitterle (Stevens)

Purpose

SERC is supporting the Undersecretary of War for Research and Engineering’s Office of Systems Engineering and Architecture on enterprise architectures (EAs) for military Systems of Systems (SoS).

Goal

Identify a pathway through which EA development and practice may evolve to meet new needs in enterprise-wide architecting for military SoS and become a construct for Government-Contractor co-creation and use.

Background

Architecture must help decision makers see how well integrated capabilities perform and how design choices affect operations and mission outcomes.

Currently, EAs for military SoS are dominantly descriptive with minimal analytical mapping capabilities, and practice is heavily influenced by tooling and a historical model-based systems engineering (MBSE) focus on artifacts and traceability. Moreover, current frameworks and the tools that implement them lack the expressivity to represent kill webs and continuous-state operation networks that are vital to mission success.

SERC advocates for a shift from descriptive artifacts to SoS architecture as a dynamic process for design exploration. SERC recommends a formal semantic ontology captured as subject-verb-object statements as a necessary foundation of an architecture model that is computable and can link the physical and functional representations of the architecture to analytical methods.

“Digital engineering will enable stakeholders to interact with digital technologies and solve problems in new and ground-breaking ways.”
– DoD Digital Engineering Strategy, 2018

SERC’s Approach

Identify priority challenges, characterize potential solutions, and develop a scalable path forward through literature review.

The SERC research team reviewed literature and frameworks to identify pain points across the current state of practice and research. In the context of military-focused SoS architecture, the team’s effort identified the highest priority challenges and devised a scalable path forward that can be achieved through a combination of Government, vendor, and research development and collaboration.

The SERC research team outlined this approach and described how it can directly enable enterprises to make various types of analytical decisions about capability selection, modernization, obsolescence, and resource allocation across all three military SoS types (kill chains, kill webs, and continuous capability networks) while showing whether these choices improve mission outcomes.

As an SoS expands, its practical distinction from EA blurs, and EA is already fragmented across strategic, technical, and governance descriptions. Furthermore, warfighting capabilities span three primary SoS types, each with unique representational needs and operational implications.

Frameworks that capture dynamic operational conditions, emergent behaviors, or shifting relational topologies are necessary but remain underdeveloped, while current EA frameworks tend to model static configurations and existing behaviors. Consequently, an EA-modeled SoS may be internally consistent by framework rules and yet structurally misrepresent real-world military governance and operational behavior.

The Three Types of Warfighting SoS

• Continuous-state network: A distributed, interconnected network of persistent infrastructure and capabilities as a continuous-state operational architecture (e.g., integrated multi-modal sensing, communications, and compute).
• Kill web: A resilient, multi-path architecture that can be dynamically orchestrated and adapted to accomplish a mission.
• Kill chain: A linear sequence, optimized as a single pathway in a kill web to execute all or part of a mission.

The SERC research team’s review of the current literature identified three key gaps:

• Gap 1: Inadequate semantic foundations. Current architecture frameworks, metamodels, and tools lack the expressive capacity to represent heterarchical SoS behaviors or to support causal reasoning.
• Gap 2: Descriptive models lack analytical power. Decision makers need to understand performance of capabilities integrated in architecture, not just architectural component structure. Current tools focus on descriptive fidelity; they do not produce inherently executable models that will show mission performance for the SoS.
• Gap 3: Architectural tempo lags operational tempo. Adversarial adaptation, funding volatility, and constituent system changes now outpace the tempo of enterprise architectural description. Automating EA will be a necessary enabler.

Based upon the findings, the team recommended the following:

• Move from descriptive to analytical, decision-centric architecting. Architectures should be machine-readable and machine-understandable, extensible, interoperable with other model types and assessment data, and quickly generated and adapted.
• Build on a formal semantic ontology captured as subject-verb-object statements. This links physical and functional representations to analytical methods – i.e., enabling descriptions to be computable – and should carry the lowest possible friction for integrating with other tools and data, including artificial intelligence.
• Adopt a scalable, integrable semantic approach covering all three warfighting SoS types. Anchoring on a semantic foundation will enable consistent definition of capabilities, roles, functions, and attributes of things and their relationships in ways that support advanced and varied approaches to analytics, e.g., directly generate network-level, queryable mathematical models to provide insight into dynamic resource allocation and adaptivity that impact operational behavior and capability portfolio investment.
• Extend EA beyond the enterprise. Develop governance representations that focus on SoS mission outcomes, examining how different governance models affect mission success.

“Having a disciplined process for architecture development in support of analytics will produce quality results, not be prone to misinterpretations, and therefore, be of high value to decision makers and mission outcomes.”
– DoD Chief Information Officer, DoDAF 2.0, Architecture Development

The SERC research team developed a recommendation for a graph-schema anchored approach utilizing Semantic Web and Linked Data standards and technologies as a bridge to facilitate achieving a unified and extensible analytical substrate. Scoped by an explicit decision taxonomy that defines what data, relationships, and performance measures each critical decision class requires, this foundation can generate the structures needed for multiple types of mathematical analysis from network-level mission resiliency to capability portfolio investment decisions.

• RDF 1.2 foundation: This newer version of the Semantic Web standard captures subject-verb-object relationships and lets properties directly attach to these. For example, conditions such as jamming state, mission phase, and condition-based authorization can attach directly to a relationship between two elements—allowing representation of concepts absent from current EA and SoS frameworks.
• Automatic consistency checking: Pairing RDF 1.2 with formal logic reasoning allows the model to be checked as it is built.
• Generated graphs: From this base, users can produce capability graphs (what each SoS type can do for a mission or scenario) and instance graphs (specific systems under specific conditions at a given time). This approach aligns with the current mission engineering paradigm, while also enables computational linking of the physical and functional representations of the architecture to analytical methods.
• Unified analytics: The single, consistent semantic foundation generates structures for multiple mathematical analyses, including tracing alternate pathways, computing feasible reconfigurations, and tracking network change over time, without separate disconnected models.
• Decision support: This enables analysis of node criticality, redundancy, and adaptivity under changing conditions, all tied to the same architecture data used to assess mission outcomes and inform investment decisions. The approach can also represent all three SoS types equally.

Looking Ahead

SERC recommends a follow-on effort with industry and Government to:

• Define the critical elements of a decision taxonomy that specify what data, relationships, and performance measures are required for key decisions with respect to more comprehensively understanding mission capabilities and performance trades in SoS, including kill webs and continuous-state networks.
• Develop the methods, processes, and foundations for the semantic base to analytical execution approach, starting with an RDF 1.2 representation of the decision architecture.
• Demonstrate the flow from semantic representation to capability graph, instance graph, and analytical graphs, using one simple exemplar per SoS type. This demonstration will refine the path forward for collaboration and transition.

“The World Wide Web connected information; the Semantic Web connected meaning; our research connects meaning to executable analysis and mission decisions.”
– Valerie Sitterle, SERC Deputy Chief Technology Officer

Follow SERC on LinkedIn for regular updates on systems engineering research.