Another new feature, Atomic batching, should help organizations that require transactional integrity across business processes, such as an online merchant that needs to make sure orders are captured even when a component such as a hard drive fails right in the middle of a transaction. Previously, developers would have to build processes, such as retry mechanisms, into their code to guarantee transactional integrity.
The CQL language, used to query and fetch data from Cassandra, has also been significantly updated. The updated version of the language, CQL3, offers a number of new calls that simplify the way applications work with data in Cassandra. Now, system information can be queried. Also the language recognizes collections as a query type. If, for instance, a single entry in a data store contains multiple email addresses, the language provides the syntax to retrieve all these email addresses.
The version of CQL will help "people wrap their minds around how to model applications of Cassandra. This meets a lot of common challenges," Ellis said.
Other new features included in Apache Cassandra version 1.2 include request tracing -- a debugging tool -- and configurable policies for disk failure. When a disk fails, an administrator can choose to stop operations entirely or continue operations on a best-effort basis. Various performance improvements and bug fixes have also been made.
Cassandra is available as open source under the Apache License version 2. Created in 1999, the Apache Software Foundation oversees about 150 volunteer-driven open source projects.