What does the Department of Defense (DoD) Chief Digital and Artificial Intelligence Office (CDAO) Data Mesh Reference Architecture (DMRA) provide?
The Department of Defense (DoD) Chief Digital and Artificial Intelligence (CDAO) Data Mesh Reference Architecture (DMRA) provides a blueprint to guide and constrain the instantiations of data mesh solution architectures.
Why is the DMRA considered as an organizational asset?
The DMRA is considered an organizational asset which provides common language for the various stakeholders, develops connection points among solutions’ implementations, supports the validation of solutions against proven Reference Architectures (RA), and encourages adherence to common patterns.
How the DMRA is developed?
The DMRA is developed from an enterprise-level perspective in support of the Department’s unified approach across data, analytics, infrastructure, and Artificial Intelligence (AI) activities; it uses a strategic approach to guide decentralized data management action across DoD, as outlined in the DoD Data, Analytics, and Al Adoption Strategy, to accelerate and scale decision advantage outcomes towards the Department’s digital transformation goals.
Which DoD Mission Areas the DMRA supports?
The DMRA supports two DoD Mission Areas: Boardroom and Battlefield. Actors in these Mission Areas, which may be human or machine, require capabilities such as those on the righthand side to the mesh.
What is the DoD definition for Reference Architecture (RA)?
Reference Architecture is an authoritative source of information about a specific subject area that guides and constrains the instantiations of multiple architectures and solutions.
What is the goal for interoperable mesh?
The goal is to have a fully interoperable mesh that enables easily discoverable and accessible data across data domains to accelerate decision advantage at speed and scale. The DMRA presents conceptual architectures of how to best orchestrate, or pull together, the DMRA elements. It does this by demonstrating that theoretical work performed to date is operationally viable versus a strongly-defined structure.
What are the benefits of the DMRA?
The DoD enterprise benefits from the DMRA in multiple ways: it achieves clarity beyond the high-level strategic work that is being performed by others on what the mesh will look like, it gives each CDAO directorate the tools to know how they will interact with it, it is rooted in tactical knowledge of best practices in the latest offerings, both internal and external in industry, this awareness is critical to optimizing the mesh design for DoD’s unique mission set, and where necessary, the DMRA guides the design of SAs.
With which CDAO strategic objectives the DMRA aligns?
The DMRA is positioned to align with the following CDAO strategic objectives: Improve Foundational Data Management through Domain Ownership & Data as a Product; Deliver Capabilities for Enterprise and Joint Warfighting Impact through Domain-Oriented Decentralization for Analytical data; Strengthen Governance and Remove Policy Barriers though Federated Computational Governance; Invest in Interoperable, Federated Infrastructure through Self-Serve Data Infrastructure Platform; Advance the Data, Analytics, and Al Ecosystem through Data Mesh Architecture; and Expand Digital Talent Management through Enhanced Technical Foundation.
What the Capability Viewpoint (CV) CV-1 Vision provides?
The Capability Viewpoint (CV) CV-1 Vision provides the strategic context for the capabilities described in the DMRA. It communicates the strategic vision for the capability areas fulfilled by the mesh and describes how the strategic vision and high-level goals and objectives should be delivered to overcome the strategic challenge on the lack of data interoperability as identified in the problem statement.
Which are the four foundational principles of a mesh?
Domain ownership mandates the domain teams to take responsibility for their data, analytical data should be composed around domains, and analytical and operational data ownership is moved to the domain teams, away from the central data team. Data as a product projects a product thinking philosophy onto analytical data, i.e., there are consumers for the data beyond the domain.
Which are the four foundational principles of a mesh? (Cont.)
The domain team is responsible for satisfying the needs of other domains by providing high-quality data. Basically, domain data should be treated as any other public API. Self-serve data infrastructure platform adopts platform thinking to data infrastructure. A dedicated data platform team provides domain-agnostic functionality, tools, and systems to build, execute, and maintain interoperable data products for all domains. Federated computational governance achieves interoperability of all data products through standardization, which is promoted through the whole data mesh by the governance group.
Which are the actors/roles within a DMRA Use Case?
Domain Team is a cross functional team of data professionals focused on developing data discoverable and easily-utilized products. Data Product Owner oversee the development and delivery of data products.
Which are the actors/roles within a DMRA Use Case? (Cont.)
Data Engineer creates and maintains DPs that serve a specific domain. Ensures that the DPs products are reliable, secure, scalable, and accessible to the data consumers. Collaborates with other data engineers and data product owners across the data mesh to share best practices and align on common standards. Data Consumer performs analysis, makes decisions, or creates products or services.
Which are the actors/roles within a DMRA Use Case? (Cont.)
Primary role is to access and explore data that is relevant, reliable, and understandable for their needs. May use various tools and methods to query, visualize, or manipulate data, depending on their skills and goals. This actor can come in many forms for various use cases. i.e. Service or Robot Accounts, ML Engineers, Data Engineers, Data Scientists or Analysts.
What does the Unique Identifier (UID) function do?
Unique Identifier (UID) function is to generate a unique sequenced UID inside the applicable domain. The capability dependencies include systematic notification that a new UID is required, ability to generate a new child UID inside of that domain for the three children listed above, and trigger to generate a unique sequential child ID recognizable by multiple domains
What the Semantic Services do?
Semantic Services: Upon receiving an element of any given language set, identify whether that element exists elsewhere in the enterprise’s known vocabularies. Notably, this includes comparing to federated dictionaries, not just a central standard dictionary. Multiple steps go into making this possible: the service must first receive the contextualized term and definition, then compare against the existing lexicons. It will either seed a new term or trigger review by the CCV governance body. Next, it will publish to the catalog and scan for additional collisions.
What is the Federated Data Catalog?
Federated Data Catalog is a reference containing metadata about DoD’s data. When well-constructed and combined with CCV, UID, and BOMs, it enables searching across all referenced data assets (including data products). The definition of ‘well-constructed’ indicates an organized, governance compliant, policy aligned listing of all data assets, recorded in the catalog with all necessary characterizations for discovery.
What is the Data and Metadata Profiles (xBOMs)?
Data and Metadata Profiles is a machine-readable (e.g., JSON, XML) listing of all the parts, subcomponents, and assemblies making up an asset. Enables dynamic management of assembly and sub-assembly, asset discovery, and insight into lineage, provenance, and pedigree. Analytic techniques to be used on [x]BOMs range the gamut from a simple search to advanced Al
What is the Policy Access Control?
Policy Access Control manages access to data products using metadata about entity actors paired with digital policy administration. The Policy Access Control service enforces the access control disposition created by using computational rule logic driven by the specified attributes. Mesh objects (data and resources) are accessible if and only if the attributes and rule logic compute a favorable access control disposition.
What does the Digital Policy Administration service do?
Digital Policy Administration manages the creation, maintenance, and auditing of digital policies. The Digital Policy Administration service provides a logic rule authoring capability, a test and evaluation environment to simulate outcomes, and a publishing service to load the Policy Access Control service with the new digital policy.
What the CV-4 Capability Dependencies Matrix describes?
The CV-4 Capability Dependencies Matrix describes the relationships among the mesh service capabilities, defining logical groupings based on the need for those elements to be integrated. The CV-4 provides a means of analyzing the dependencies among mesh service capabilities. The groupings of capabilities are logical with the purpose to guide enterprise service management.