Graph databases

The main standard in this area is Apache TinkerPop v3 , with an ecosystem centered around the JVM.

• Neo4j
  • License: AGPL + commercial
  • Most popular graph DB
• Amazon Neptune
  • A property graph: supports Apache TinkerPop Gremlin
  • Also a triple store: supports RDF 1.1 and SPARQL 1.1
  • Rumored to be based on Blazegraph , which powers Wikidata’s SPARQL endpoint
• JanusGraph (GitHub )
  • License: Apache 2
  • Uses existing DBs as backends
  • Fork of TitanDB , which is abandoned
  • Basis for IBM Graph, now retired
• OrientDB
  • License: Apache 2
  • Multi-model including graph DB
• Cayley (GitHub , HN )
  • “Inspired by graph database behind Freebase and Google’s Knowledge Graph”
  • Uses existing DBs as backends
• Dgraph (GitHub )
  • License: Apache 2
  • Written in Go with official clients for Go, Java, JavaScript/TypeScript, and Python (plus universal gRPC and HTTP protocols)
  • Data entry via RDF triples (primary) and JSON (converted to RDF)
  • Custom query language GraphQL+- inspired by Facebook’s GraphQL
  • Differences between Dgraph and Cayley

• Angles & Gutierrez, 2008: Survey of graph database models (doi)
  • Angles & Gutierrez, 2005: Querying RDF data from a graph database perspective (doi)
  • Angles, 2012: A comparison of current graph database models (doi)
• Rodriguez & Neubauer, 2010: Constructions from dots and lines (doi, arxiv)
  • Rodriguez & Neubauer, 2010: The graph traversal pattern (doi, arxiv)
  • Rodriguez, 2015: The Gremlin graph traversal machine and language (doi, arxiv)
• Wood, 2012: Query languages for graph databases (doi)
• Barcelo et al, 2013: Schema mappings and data exchange for graph databases (doi)
• Besta et al, 2020: Demystifying graph databases: Analysis and taxonomy of data organization, system designs, and graph queries (arxiv)
  • Figures 4 and 5 show translation from labeled property graph to RDF triples
  • Figure 7 is a nice overview of different kinds of graph database systems