Uploaded byGalder Zamarreño
Data Grids vs Databases
This document summarizes a presentation about data grids versus databases. It discusses how data grids provide extremely fast in-memory access to distributed data and can be used for caching to improve database performance. While data grids offer benefits like scalability, their use requires a different programming model than databases. They may replace databases for some use cases like analytics but databases will remain important for their maturity and existing implementations. Data grids are best viewed as complementing rather than replacing databases.
Data Grids vs Databases
- 1.
- 2.
- 3. Data Grids vs Databases Galder Zamarreño Senior Software Engineer Red Hat, Inc 3rd October 2011, Soft Shake Thursday, November 3, 11
- 4. Galder Zamarreño • R&D Engineer, Red Hat Inc. • Infinispan developer • 5+ years exp. with distributed data systems • Twitter: @galderz • Blog: zamarreno.com Thursday, November 3, 11
- 5. Agenda • Why do we need Data Grids? • What are exactly In-memory Data Grids? • Data Grids + Databases • Data Grids without a Database • Can Data Grids replace Databases? Thursday, November 3, 11
- 6. Traditionally... Store everything in a DB! Thursday, November 3, 11
- 7. Modern requirements DBs not particularly good at horizontal scaling... Thursday, November 3, 11
- 8. One size doesn’tfill all! DBs are not bad, but they’re not the solution to every problem either Thursday, November 3, 11
- 9.
- 10. Data Grids arenot new Mainstream traction only recent: vertical scaling needs, cheaper memory... and cloud! Thursday, November 3, 11
- 11. Who’s offering Data Grids? Thursday, November 3, 11
- 12. The Players • Open Source: • Infinispan, EhCache, Hazelcast... • Commercial: • Oracle Coherence, Gigaspaces, Gemfire, IBM eXtreme Scale Thursday, November 3, 11
- 13. But, what are In-memory DGs? Thursday, November 3, 11
- 14. Definition In-memory data structures that offer extremely fast access to data Thursday, November 3, 11
- 15. Maps are popular! Normally come with a Map-like API, but often come with alternatives Thursday, November 3, 11
- 16. Data distribution Store data in a subset of the grid to provide failover while being able to scale up! Thursday, November 3, 11
- 17. With failure inmind Suitable for commodity hardware because they can handle failure Thursday, November 3, 11
- 18. Elastic Remain available during topology changes Thursday, November 3, 11
- 19. Durability More durability achieved flushing to a persistent store Thursday, November 3, 11
- 20. Access patterns Embedded (client and DG in same VM) or Remote (just like DBs) Thursday, November 3, 11
- 21. ACID or BASE Transactions or Eventual Consistency? Thursday, November 3, 11
- 22. DGs + DBs? Thursday,November 3, 11
- 23. Caching! Use Data Grids as caches to enhance Database access performance! Thursday, November 3, 11
- 24. Can a DataGrid replace a DB? Thursday, November 3, 11
- 25. Reiterating benefits Speed, scalability, cloud- friendliness...etc Thursday, November 3, 11
- 26. What are theData Grid challenges? Thursday, November 3, 11
- 27. Access patterns Migrating from SQL to Map or alternative APIs not easy Thursday, November 3, 11
- 28. Skill set Different skill set: OO programmer vs SQL Thursday, November 3, 11
- 29. Application data layer Data layer to take data collocation into account and do more validation (less strict schema) Thursday, November 3, 11
- 30. E.g. with aDB... Thursday, November 3, 11
- 31. Same with Infinispan Thursday,November 3, 11
- 32. Map/Reduce in detail Thursday,November 3, 11
- 33. Technology to bridge gap? Thursday, November 3, 11
- 34. What about JPA? Hibernate OGM (Object/Grid Mapper) uses JPA to store in DGs as opposed to DBs Thursday, November 3, 11
- 35. Most frequent use cases for DGs? Thursday, November 3, 11
- 36. Use cases • Analytic systems, i.e. financial/trading apps • XTP • Event driven apps, i.e. CEP • Clustering toolkit Thursday, November 3, 11
- 37. Do I seeDGs as DB replacements? Thursday, November 3, 11
- 38. DBs are hereto stay! No. DBs are proven, mature, well understood plus, there are millions of systems out there! Thursday, November 3, 11
- 39. One size doesn’tfill all! DBs are not a universal data storage system any more Thursday, November 3, 11
- 40. Consider Data Grids For their speed, capabilities as data store, and cloud friendliness Thursday, November 3, 11
- 41. Still some wayto go More deployments and standardization (JSR-107, JSR-347) Thursday, November 3, 11
- 42. Questions infinispan.org - @infinispan speakerrate.com/galder Thursday, November 3, 11