This wiki is deprecated, please use this one instead
- Cassandra is designed to be always available. Writes never fail. Two read paths are available: high-performance "weak" reads and quorum reads. See ThriftInterface.
- Cassandra has a rich data model allowing efficient use for many applications beyond simple key/value.
- Data is automatically replicated to multiple nodes for fault-tolerance. There is support for implementing strategies that replicate across multiple data centers.
- Elasticity: new nodes can be added to a running cluster while minimizing disruption to existing data.
- Consistency: Cassandra follows the "eventually consistent" model but includes sophisticated features such as Hinted Handoff and Read Repair to minimize inconsistency windows.
- Reads and writes in Cassandra are guaranteed to be atomic within a single ColumnFamily.
- Support for versioning and conflict resolution (with inbuilt policies like "last update wins").
Cassandra is not a relational database. You cannot run SQL queries in Cassandra or write ACID transactions. However, if your application needs high-performance, elastic, scalable, and structured storage try Cassandra!
Many of the Cassandra developers and community members hang out on #cassandra on freenode irc: http://www.mibbit.com/?server=irc.freenode.net&channel=%23cassandra&nick=yournamehere