XFSC: A Catalogue of Trustable Semantic Metadata for Data Services and Providers
Benedikt T. Arnold, Khalil Baydoun, Diego Collarana, Sebastian Duda, Christina Gillmann, Ahmad Hemid, Philipp Hertweck, Paul Moosmann, Denis Sukhoroslov, Christoph Lange
TL;DR
The paper addresses the challenge of discovering trustworthy data resources in decentralized dataspaces by introducing the XFSC Catalogue, a trustable semantic metadata directory aligned with Gaia-X. It presents a full stack consisting of Self-Description lifecycle management, schema-based consistency enforcement, a rigorous three-stage verification (syntax, cryptography, and semantic validation), and a scalable queryable graph store using openCypher on Neo4j with neosemantics. It further details integration with the Eclipse Dataspace Components Connector via a Broker EDC, enabling real-world use in the German Culture Dataspace, and provides a complete implementation in Java with a REST API, RBAC, and robust testing. The work demonstrates the Catalogue’s practicality, scalability, and trustworthiness, highlights current limitations, and outlines future directions including newer Gaia-X spec compatibility, deeper EDC integration, and broader federation adoption."
Abstract
In dataspaces, federation services facilitate key functions such as enabling participating organizations to establish mutual trust and assisting them in discovering data and services available for consumption. Discovery is enabled by a catalogue, where participants publish metadata describing themselves and their data and service offerings as Verifiable Presentations (VPs), such that other participants may query them. This paper presents the Eclipse Cross Federation Services Components (XFSC) Catalogue, which originated as a catalogue reference implementation for the Gaia-X federated cloud service architecture but is also generally applicable to metadata required to be trustable. This implementation provides basic lifecycle management for DCAT-style metadata records and schemas. It validates submitted VPs for their cryptographic integrity and trustability, and for their conformance to an extensible collection of semantic schemas. The claims in the latest versions of valid VP submissions are extracted into a searchable graph database. The implementation scales to large numbers of records and is secure by design. Filling the catalogue with content in a maintainable way requires bindings towards where data and service offerings are coming from: connectors that expose resources hosted in an organization's IT infrastructure towards the dataspace. We demonstrate the integration of our catalogue with the widely used Eclipse Dataspace Components Connector, enabling real-world use cases of the German Culture Dataspace. In addition, we discuss potential extensions and upcoming integrations of the catalogue.