The year of the graph: do you really need a graph database? How do you choose one?
- 1. THEYEAR OFTHE GRAPH: DOYOU REALLY NEED A GRAPH DATABASE? HOW DOYOU CHOOSE ONE? The year of the graph: do you really need a graph database?The year of the graph: do you really need a graph database?
- 2. ABOUT ME Working with data since 1992 Graph DBs since 2005 Databases Modeling Research Analysis Consulting Entrepreneurship Journalism
- 3. THE PROBLEMWITH RELATIONAL DBS Not good at relations! * Unintuitive model * Hard to query * Does not scale
- 4. GRAPH USE CASES: OPERATIONAL APPS Image: Neo4j
- 5. GRAPH USE CASES: OPERATIONAL APPS Smart Home - IoT * LeadingTelco in the Nordics * 1,5 Million Homes * Real-time processing
- 6. GRAPH USE CASES: ANALYTICS Image: Stanley Wang
- 7. GRAPH USE CASES: ANALYTICS Drug Discovery * Leading Pharma * Data on genes, proteins, etc * Identification of causal relationships
- 8. GRAPH USE CASES: DATA INTEGRATION Image: Ontotext
- 9. GRAPH USE CASES: DATA INTEGRATION Knowledge Graph for Search * Leading Retailer in DACH * 200Million+ MAU, 300K+ search requests * Improve coverage, response time, bottomline
- 10. GRAPH USE CASES: MACHINE LEARNING Image: Oracle
- 11. GRAPH USE CASES: MACHINE LEARNING Anti-Fraud in real-time * LeadingTelco in China * 600 Million Users * Compliance, trust
- 12. KNOWLEDGE GRAPHS EVERYWHERE, GRAPH DATABASES ASTHE FOUNDATION
- 13. GRAPH DATABASES ARE BOOMING.. SO HOW DOYOU CHOOSE ONE? Existing research is: * Outdated * Shallow * Expensive * Marketing oriented
- 14. EVALUATING GRAPH DATABASES: HTTP://YEAROFTHEGRAPH.XYZ Premises: * Always up to date * Holistic evaluation *Value for money * Hands-on Why me: * Hands-on since 2005 *Top-tier analyst since 2013 * Independent Free Newsletter!
- 15. WHAT GRAPH DATABASES ARE NOT: ANALYTICAL &VIZ FRAMEWORKS,THIN GRAPH LAYERS
- 16. REAL GRAPH DATABASES GO ALLTHEWAY Operational vs. Analytical * Fully-fledged graph API * Operations & Analytics * Future-proof, integrated Native vs. Non-native *Designed as a graph database * Storing data in a native format * Optimized for graph
- 17. GRAPH DATABASETYPES: LPG (LABELED PROPERTY GRAPH) * Non-standard format, query * Poor schema support * Interoperability * Flexible, generic * Fast traversals * Scalability
- 18. LPG GRAPH DATABASE USE CASES Operational applications Graph Traversals Graph Algorithms
- 19. GRAPH DATABASETYPES: RDF (RESOURCE DESCRIPTION FRAMEWORK) * W3C: Interoperability, Maturity * Rich & Flexible Schema * Semantics, Rules, Reasoning * Performance * Scalability * Complexity
- 20. RDF GRAPH DATABASE USE CASES Data Integration Knowledge Graph AI
- 21. GRAPH DATABASETYPES: MULTI-MODEL (DOCUMENT, KEY-VALUE) * Flexibility *Tooling * May not be optimized for graph * May have to move data around * Lock-in (cloud vendors)
- 22. EVALUATION CRITERIA: HOLISTIC, DATA-DRIVEN Application Development * Engine & API * Data Model * Query Language DevOps * Interoperability * Deployment & Configuration Advanced Analytics *Advanced Graph *Visualization Vendor Credentials & Support *Vendor Credentials * Support & Community
- 23. EVALUATION CRITERIA: PERFORMANCE & COST * Not all vendors participate in benchmarks * Most benchmarks are not done by 3rd party * Complex, demanding exercise * Hard to compare LPG vs RDF * Not all vendors have public pricing & license info * Considered, not included as data points * Informed judgement needs experience, hands-on
- 24. TAKEAWAY POINTS * Graph DBs are here to stay * Different Graph DB models LPG RDF Multi-model * Each best suited to different use cases * Evaluation is hard!
- 25. EVALUATING GRAPH DATABASES IS HARD.. BUT SOMEBODY HASTO DO IT Impressive work. I’m not aware of another source that is as comprehensive as this one. JONATHAN LACEFIELD, SENIOR DIRECTOR OF PRODUCT MANAGEMENT, DATASTAX ENTERPRISE SERVER We did not have the time, resources, or expertise to evaluate all options properly. If we did, our choices would have been different. APPLICATIONARCHITECT
- 26. EVALUATING GRAPH DATABASES: JUST DOTHE MATH * 30+ options * Costs time and money * Requires expertise * Lack of proper evaluation -> Sub-optimal decisions -> Report cost: 1 Day ofTop-tier consultant Access to updates, consulting