Database
Amazon Neptune
Amazon Neptune is a managed graph database supporting property-graph and RDF workloads. A GenAI application can combine graph relationships with retrieval to answer connected, provenance-sensitive questions.
Key points
- Gremlin and openCypher address property graphs while SPARQL addresses RDF
- Read replicas scale graph-query traffic
- Neptune Analytics and database capabilities should not be treated as interchangeable deployment modes
When to use it
- Retrieve connected entities for GraphRAG
- Trace policy, asset, and owner relationships for an agent
Exam tips
- Choose Neptune when traversals are central; a vector store remains better for semantic similarity alone
- Enforce authorization during traversal because filtering only the final answer can expose restricted relationships