Standards Interop

The OSS reference implementation of the IDSA Dataspace Protocol — handles inter-organizational connectors, policy negotiation, and data transfer.

edc-source / edc-sink adapters are planned as a separate repository implementing the OSS Source / Sink extension-point contracts (extension adapters live outside the OSS core), so EDC-negotiated data can be piped through provin to attach provenance VCs.

“How parties connect” (Dataspace Protocol) and “what is guaranteed about the data” (provin) split cleanly. Combine them to add a provenance layer on top of any IDSA / GAIA-X data space.

An industry-standard specification for pipeline lineage under the Linux Foundation, with native support across Airflow, dbt, Spark, and other major data tools.

provin’s pipeline/observer is designed as a hook for processing-completion events. Implement an OpenLineage emitter on top of it to forward provin pipeline activity as OpenLineage Run Events to any compatible lineage backend.

Not yet shipped as an official adapter, but the ProcessObserver interface is built to be extended by the community.

The reference OpenLineage server, providing lineage visualization and history.

Connecting provin via an OpenLineage emitter exposes pipeline provenance as a time-series graph in the Marquez UI.

Two of the most widely used OSS metadata management platforms — unifying data assets, schemas, and lineage across an organization.

Register provin VCs to DataHub / Atlas metastores via pipeline/observer to surface provenance-verified data assets in your external catalog.

An open protocol for decentralized data pipelines (reference implementation: kamu). It stores data and metadata as cryptographic ledgers and verifies query results through deterministic recomputation.

A kindred approach in the same problem space. ODF verifies the result — by ledgering the data itself and making transformations reproducible; provin verifies the attribution — by signing boundary commitments. Tolerating non-deterministic stages (model inference, human judgment) is provin’s side of the split.

Interop is implementable at the Source / Sink extension points, treating ODF datasets as pipeline inputs and outputs.