How eclipse-edc/federatedcatalog works
Project type and purpose
• This is a standalone backend microservice—“EDC Federated Catalog”—built on the Eclipse Data Space Connector (EDC) runtime.
• It periodically crawls a list of remote data-space participants, fetches their catalogs, transforms and caches them locally, and exposes a unified catalog API.
Key modules (from the file tree)
• spi/…-spi: Java Service Provider Interfaces defining crawler and federated‐catalog contracts
• core/…: core implementations (common utilities, crawler engine, federated-catalog logic, JSON transformers)
• extensions/api/federated-catalog-api: HTTP API extension (REST controllers)
• extensions/store/sql/…: optional SQL-backed implementations for cache and directory
• launchers: Docker/Java process entry points (in-memory “mocked” or DCP-secured identity)
• system-tests: integration and end-to-end tests
Architectural overview
• At runtime, the EDC Connector bootstrap loads all ServiceExtension implementations (from core, extensions, sql modules, etc.) into a single-process container.
• A Directory component holds the list of participant endpoints (in-memory or SQL).
• An ExecutionManager schedules crawler tasks at regular intervals.
• Each CrawlerAction fetches a participant’s catalog over HTTP (or via DCP-secured channel), hands off the raw JSON to a JsonObjectToCatalogTransformer, and then invokes the FederatedCatalogCache to store/update entries.
• The FederatedCatalogCache sits on top of an in-memory map or a SQL store, depending on which store extension is active.
• The FederatedCatalogApiController exposes /api/catalog over HTTP; when a client query arrives, it goes through QueryService → FederatedCatalogCache → JSON serializer → HTTP response.
• An external Identity Provider (mocked or DCP) is used for authentication/authorization on crawl requests and API calls.