Scaling Identification Solutions Using Graph DB
summary
Target replaces a vendor relational identity product with a graph architecture using JanusGraph + Cassandra (after experimenting with Neo4j), ingesting anonymized feeds and matching records to produce a unified view of guests. The system runs at ~1 billion nodes and ~3 billion edges, supports real‑time personalization, targeted ads and churn detection, cuts licensing and scaling costs, and uses techniques like avoiding supernodes, smart indexing and eventual‑consistency guards.
AI-written summary · about
chapters (10)
- 0:00Introduction to scalable systems
- 3:18Key factors for scalability
- 9:10Why graph databases fit
- 13:24Graph database basics
- 21:18Real-world graph database use cases
- 28:41Target’s graph database journey
- 29:01Guest identity resolution at Target
- 33:59Building in-house solution
- 43:16Engineering lessons learned
- 45:27Conclusion and future outlook
related talks
36mIntroduction to JanusGraph Database
36mIntroduction to JanusGraph Database
25mBuilding the World’s Largest Identity Graph without a Graph DB
30mLarge Scale Graph Solutions: Use-cases and Lessons Learnt
21mFrom Data to Knowledge to Action - The Graph Intelligence Platform
30mPerformant GraphRAG Improving Neo4j to Drive Autonomous AI ...
19mNuGraph: GraphDB as a Cloud Service Built Upon JanusGraph and FoundationDB
37mGraph Computing at Scale