Download as PDF, PPTX
Uploaded byApigee | Google Cloud
Building a Mobile Data Platform with Cassandra - Apigee Under the Hood (Webcast)
The document discusses Usergrid, an open source mobile backend platform built on Apache Cassandra. It provides capabilities like API management, analytics, and tools. Usergrid allows building mobile and rich client apps without needing a web stack. It highlights key features like being platform agnostic, flexible data modeling, and multi-tenancy using virtual keyspaces in Cassandra. The document also discusses how Usergrid implements shared schemas and keyspaces in Cassandra to provide isolation and scale for multiple teams and applications.
More Related Content
Similar to Building a Mobile Data Platform with Cassandra - Apigee Under the Hood (Webcast)
![Support, Monitoring, Continuous Improvement & Scaling Agentic Automation [3/3]](https://cdn.slidesharecdn.com/ss_thumbnails/agenticcommunityseries-day3-cfd-251120170304-ddef8112-thumbnail.jpg?width=640&height=640&fit=bounds)
Building a Mobile Data Platform with Cassandra - Apigee Under the Hood (Webcast)
- 1. Usergrid Open Source Mobile Backend on Cassandra
- 2. Hi I m EdAnuff
- 3.
- 4. Usergrid Backend-as-a- Service & Open Source Mobileand Rich client apps
- 5. App, Cloud, &HTTP
- 6. No Web Stack Page Templates MVC Logic Database
- 7. Just Apps andAPIs
- 8. Key Capabilities A platformnot a single app Flexible data model Easy to use from any language
- 9.
- 10. Open Source (APLv2) Builton Hector Multi-tenant (Virtual Keyspaces) JSON and full-text indexing Rich permissions model REST API
- 11.
- 12. Lots of Teams... Team Team Team
- 13. Lots of Apps... App App App Team
- 14. And App Users... App Team
- 15. And data and devices App Team
- 16. Multi-tenancy at scale App App App App App App Team Team Team
- 17. Schema Separate Database Shared Database Shared - Scalability + Scalability + Isolation - No Isolation + Simple - Complicated Separate Schema - Scalability - Scalability + Isolation - No Isolation + Not Complicated - Complicated
- 18. CREATE DATABASE ALTER TABLE
- 19. Sharding SQL istoo much work. Oracle is too much money.
- 20. What about NoSQL? MostNoSQL works at the wrong abstraction level.
- 21. Does Cassandra provide anymore options? Sort of
- 22. Schema Separate Keyspace Shared Keyspace Shared Expensive Interesting Separate Schema Very Unwieldy Expensive
- 23.
- 24. Users = { "4e3c0423-…" : { username: "edanuff", email: "[email protected]", … : … }, "e5d61f2b-…" : { username: "jdoe", email: "[email protected]", … : … } }
- 25. Users = { "4e3c0423-…" : { username: "edanuff", email: "[email protected]", … : … }, "e5d61f2b-…" : { username: "jdoe", email: "[email protected]", … : … } }
- 26. Conventional Row KeysIn Single Keyspace Row UUID Row UUID Row UUID Row UUID Row UUID Row UUID
- 27. Multi-tenant Row KeysIn Shared Keyspace Tenant ID Row UUID Tenant ID Row UUID Tenant ID Row UUID Tenant ID Row UUID Tenant ID Row UUID Tenant ID Row UUID
- 28. Wonderful, more things forme to build myself in the app tier... Hold that thought...
- 29. Shared Schema/ Shared Keyspaceis Virtual Keyspaces Part of Hector Client
- 30. public static <E>Keyspace createVirtualKeyspace( String keyspace, E keyPrefix, Serializer<E> kps, Cluster cluster);
- 31. Hector hides theTenant Prefixes Prepend Tenant UUID Hector Cassandra Remove Tenant UUID
- 32. What about indexes? Hectordoesn t filter results Don t add a tenant column App-tier CF indexes preferable
- 33. What about CQL? Seeprevious slide for index limitations Hector doesn t currently support Virtual Keyspaces via CQL I don t actually ever use CQL
- 34. So what elseis in there?
- 35. JSON Full JSON values, fullyindexed, using Jackson and stuff from Lucene
- 36. Permissions Fine grained, per user,built on Shiro
- 37. REST Full REST API,built on Jersey
- 38.
- 39. Mongo Compatible API Experimental Mongo APIemulation via Netty Mongo binary wire
- 40.
- 41. Use the ideas Usethe code Use the cloud
- 42.