Implementing a Graph Database based on PostSQL
summary
Azen Graph implements a disk‑backed, OLTP graph engine on PostgreSQL, shipping v0.9 that supports 70% of openCypher, JDBC/ODBC/Python clients and a web visualizer. Vertices use two‑column tables, edges four columns and current B‑tree composite indexes make storage roughly 2× the size of Neo4j; a new single‑column, pointer‑based index (storing edge HIDs) aims for index‑only traversals and cheaper updates. Property values live as JSON (notably large), LDBC benchmarking is underway and distribution plans target Postgres‑XL.
AI-written summary · about
chapters (18)
- 0:00Intro
- 0:58What is the Graph Database?
- 3:12Cypher Query Language
- 3:53Cypher Query Example
- 5:51GraphDB to PostgreSQL Case
- 7:17Graph Database and Hybrid Database
- 8:31Why We Choose PostgreSQL?
- 10:23Challenges
- 12:46Two Graph Databases
- 13:32Graph Storage - Neo4j
- 14:49Graph Storage - Titan (DSE Graph)
- 16:53Index Problems
- 18:31New Indexes for Graph Traversal
- 19:42Graph Storage - AgensGraph
- 20:30Current Status
- 21:18Tadpole for Agens Graph
- 22:14Future Roadmap
- 23:05Join Us
related talks
49mModern Graph Databases
41mGraphGen: Conducting Graph Analytics over Relational Databases
36mIntroduction to JanusGraph Database
30mAerospike Graph: Massively Scalable, High Performance
22mQuerying & Visualizing Graphs in Postgres with Apache AGE
01h 00mWho’s Afraid of Graphs?
29mSo... Graph Databases
27mGraphOne: A Data Store for Real-time Analytics on Evolving Graphs