Cassandra is rapidly heading towards 2.1 release, with 2.1.0-beta1 already available for evaluation. Incubated and open sourced by Facebook, it is currently used by Reddit, Netflix, Ebay, Twitter and others. Supported by DataStax, Cassandra is expanding its reach towards the enterprise world.
Cassandra version 2.0 delivered lightweight transactions, triggers, atomic batch guarantees for a large set of prepared statements and performance improvements. The Cassandra Query Language used in addition to the native drivers is based off SQL in the sense that data is put in tables containing rows of columns. Version 2.1 adds indexes on collections and User Defined Functions.
Cassandra 2.1 also uses cardinality estimation for data files compaction. The algorithm results in reducing MemTable memory usage by 85% and improving write performance by around 50%. The algorithm used is HyperLogLog, originally developed and open sourced by AddThis.
DataStax, as a contributor to the Apache Cassandra project and boasting a lineup of enterprise clients is one of the companies trying to bring Cassandra to a larger audience. Chief Evangelist for DataStax, Patrick McFadin published an article throwing the glove on MongoDB on scaling issues. To reinforce his point, Mr. McFadin included testimonials from several Cassandra users including no other than Chris Kalantzis, Cloud Database Engineering Manager at Netflix.
DataStax recently announced a partner network program. Partnering up with Accenture means that its clients will have better access to Cassandra technology. Google Cloud Platform, HP Moonshot and GoGrid on the other hand will expand the available platforms and IaaS on which DataStax and Cassandra can be deployed on.
Along with expanding its reach in the enterprise through alliances, DataStax is entering the in-memory database sector by claiming up to 100 times better in-memory performance with its commercial offering, DataStax Enterprise 4. DataStax enterprise edition also adds enterprise search with Solr, integration with Hadoop, visual management and support. DataStax or not, Cassandra is gaining enterprise adoption that can not be ignored.