Glossary

Attribute-based access control. ShrikeDB policies can match principal, resource, and environment attributes.

Graph query language used by ShrikeDB’s main graph and multimodal query entry point.

Durable storage unit. Its payload represents a particular data modality, such as a graph or vector record.

Queryable graph index derived from durable graph fragments and journals.

High availability. Cluster services coordinate membership, placement, replication, and repair.

Hierarchical Navigable Small World graph. ShrikeDB uses HNSW-backed vector search.

A query that combines two or more modalities, commonly vector candidate generation, graph expansion, and KV hydration.

Computes supported inferred facts from asserted ontology data.

Trace format used by ShrikeDB benchmark and admin trace outputs.

Versioned graph snapshot used by readers while writers continue ingesting and publishing newer state.

Role-based access control. ShrikeDB roles resolve to permissions over tenant, KV, graph, or vector resources.

Resource Description Framework. The semantic reasoning path ingests RDF/OWL ontology data.

A database storage file recorded in a manifest.

RDF graph query language supported for covered graph-pattern query shapes.

Structured materialization output consumed directly by the reasoned graph index to avoid rebuilding from flat triples.