Blockchain for Trustworthy Publication and Integration of Linked Open Data

Abstract

The timely, traceable and provenance-aware publication of Linked Open Data (LOD) is crucial for its success and to fulfill the vision of a global, decentralized, and machine-readable database of knowledge. Yet, the access to LOD is still fragmented and mainly centralized aggregations are being used, relying on complex harvesting mechanisms. As a remedy, we propose a blockchain-based approach enabling an integrated, traceable, and timely view on LOD. We use a blockchain to meet the organizational requirements of publishing LOD in a decentralized fashion while still supporting the sovereignty of the data providers and supporting provenance and proper integration into a harmonized knowledge graph. We present an approach and an implemented system that fulfills the requirements regarding volume and throughput and can be used as the foundation for practical deployments. We use Linked Open Government Data (LOGD) as our case study to demonstrate the feasibility of our approach. We developed a prototype to address the specific requirements of LOGD publication and apply the Practical Byzantine Fault Tolerance algorithm at its core to enable a robust state replication.