A FRAMEWORK FOR SEMANTIC AUDIT LOGGING AND VISUALIZATION OF A CTI KNOWLEDGE GRAPH

This thesis extends a decentralized Cyber Threat Intelligence (CTI) platform based on Hyperledger Fabric (HLF) with a Semantic Audit Logging framework that makes on-chain decisions and agent interactions observable and traceable. A formal model defines a uniform schema for audit records, aligned with a standard provenance vocabulary and linking every relevant event to the underlying knowledge graph through ontological references. The blockchain layer is extended to emit such events natively within the transactions that produce them, while a dedicated middleware consumes them through the fabconnect connector of Hyperledger FireFly, indexes them in an embedded SQLite store, and exposes a domain-oriented REST interface. On top of this middleware, a web dashboard supports the analyst in investigating system decisions, exploring entity histories, and visualizing the knowledge graph. The evaluation measures the post-commit latency of audit-record availability, the scalability of the knowledge graph and of the audit pipeline, and the loading performance of the interface, showing that, in the configuration examined, observability is obtained at a limited cost to the underlying system.