Federation Services (aka FedSrv) is Azure Microservices that creates a Data Mesh by combining an Ontological Knowledge Graph with deep API Integrations, which are Plugin-based for easy reuse.
Medium-to-Large Organizations, regardless of industry vertical, now have more APIs and data sources than ever. But to truly unlock these, they must all be “tamed”: auditable identity across a diverse array of API transaction processes must be preserved, groups of data schemas must be translated into meaningful Data Products, and so on. Furthermore: this must all be done without creating new copies of existing data or having to re-implement existing data systems.
Another distinct challenge is the ability to present the diverse array of APIs and Data Sources in a consistent way, so that massive amounts of data can be consumed by GenAI, LLM, and ML systems for the purposes of Training.
Our Unique Approach:
Avanade has created a set of lightweight, flexible Microservices called Federation Service, aka “FedSrv”. It's extremely flexible nature allows it to work with any existing integration systems such as Azure API Management, Mulesoft API Management, Microsoft Fabric, and Microsoft Azure Data Fabric.
FedSrv fundamentally consists of two parts:
1) API integration via “API Plugins”. This enables any number of APIs and Data Systems to be seamlessly integrated, and accessed via a completely consistent, OData Compliant, REST control plane.
2) Data Mesh/Knowledge Graph creation. This enables organizations to create a model that describes, in a detailed fashion, their Business Logic and Processes. This has the collateral benefit of providing a highly consistent means of training GenAI and ML systems in Business Logic and Processes.
Also, of course, these two distinct parts can be used together: a Knowledge Graph can be created, and each “type” in the Knowledge Graph can be backed by a REST API Endpoint
Key Outcomes:
Seamless integration and federation of a diverse variety of API and Data Systems. A true Knowledge Graph and Data Products. There are several specific offerings available:
1) “FedSrv for API Integration”: FedSrv Implementation for API and Data System Integration with Existing Apps and Systems.
2) “FedSrv for AI Training”: FedSrv Implementation to target Mass Data Ingestion to existing GenAI and ML Systems.
3) “FedSrv for Data Mesh”: FedSrv Implementation to enter into Data Mesh, Data Product, and Knowledge Graph Technology.
4) “FedSrv for AI Cognition”: FedSrv Implementation towards using GenAI and ML Systems to understand Organization’s Business Rules.
Deliverables:
Specific Deliverables depend on the specific Offering being implemented:
1) “FedSrv for API Integration”: A consistent Odata-compliant REST control plane, that can be handle fully auditable E2E CRUD lifecycles for all backing data, as well as the ability to see record-level Versioning.
2) “FedSrv for AI Training”: Easy consumption of a diverse array of an organization’s existing data by existing GenAI and ML Systems, providing much deeper insight into this data.
3) “FedSrv for Data Mesh”: An Ontological Knowledge Graph, complete with REST API Endpoints, that can form the basis for a Data Mesh that delivers Data Products.
4) “FedSrv for AI Cognition”: Easy consumption of an Organization’s Business Rules and Logic via GenAI and ML Systems. The basic ability to use Natural Language queries to ask meaningful questions about Business Rules and Logic.S
*Pricing will vary based on the scope of the engagement.