Nearly five months to the day after the release of the last major iteration, Cassandra 3.0 is rolling out into general availability with a couple of major improvements promising to help users handle their unstructured data much more efficiently. What they may lack in number, the project’s backers hope the additions will be able to make up in scope.

The main highlight is a reworked version of the storage framework that Cassandra employs to organize incoming information into a consistent catalog that can be reliably accessed and modified over time. The original implementation was a relic from the database’s early days as an internal project at Facebook Inc. that very much reflected the social networking giant’s relatively limited requirements at the time, first and foremost in the efficiency department.

Cassandra until now had to redefine where a particular row of data cells starts and ends before a bulk access attempt, incurring a small amount of additional overhead every time. That could add up to significant delays in large deployments, especially when running complicated queries that operate on multiple rows at a time. The project’s backers claim that eliminating this snag along with some of the more minor issues that plagued the database’s storage framework has made it possible to cut capacity requirements by an average of 50 percent in the current iteration.

Users will be able to take advantage of the increased efficiency using the other new feature introduced as part of Cassandra 3.0, a variation of the “Materialized View” capability in Oracle Corp.’s rivaling relational database that reduces the amount of effort involved in modifying large swaths of information. It’s also much faster than the manual equivalent. Robin Schumacher, the head of products at DataStax Inc., a startup working to commercialize the project, said in an interview ahead of today’s launch that the functionality was shown to deliver a speed improvement of between 50 and 80 percent during internal tests.

Image via tookapic