Skip to content

Chapter 2: Graph Databases

Chapter Overview

  • What are graph databases?
  • What problems do graph databases solve?
  • How are graph databases different from relational databases?

What are graph databases?

  • Vertices and edges
  • Both may have attributes
  • A universal structure for representing knowledge
  • Many implementations
  • RDF
  • LPG
  • Neo4j Championed the native LPG data model
  • Pointer Hopping is roughly 1,000 the performance of RDBMS JOIN operation
  • TigerGraph championed the distributed native LPG model
  • No more concerns about scale-out graphs
  • Scale out over 100 nodes allows a 100x performance speedup
  • Intel and other companies working on 1,000x performance speedup of native labeled property graphs
  • Combined, these three technologies offer a 10,000,000x performance speedup
  • Key limitations are the cost of integrating data

What problems do graph databases solve?

  • Performance on edge traversals
  • Work well with highly connected data sets
  • Excellent at modeling complex Relationships
  • Precise models are more shareable across multiple business units

How are graph databases different from relational databases?

Graph timelines

  • The seven bridges of Konigsberg
  • The Samenatic Web - 2001
  • RDF
  • SPARQL
  • Neo4j - The LPG data model