With interest in the open source NoSQL Cassandra database as platform for highly distributed applications picking up, a race is on to provide cloud and hosting services for it. The platform is starting to gain significant traction at the higher end of the enterprise IT market.
At the Cassandra Summit 2015 in Santa Clara, California, Microsoft announced it will make an enterprise edition of Cassandra by a company called DataStax available as a service on its Azure cloud. That move comes a day after Rackspace announced that it will provide a managed instance of Cassandra by DataStax as part of its service portfolio.
DataStax claims to have more than 1,000 customers running its Cassandra implementation, known as DataStax Enterprise (DSE). Cassandra itself was designed from the ground up to simultaneously support both transaction processing and analytics applications. That’s appealing to IT organizations that traditionally had to bear the cost and expense of supporting separate databases for transaction processing and analytics applications.
Because of the capability, Cassandra is currently being widely tested as an alternative to traditional relational database systems from Oracle and IBM. To make NoSQL Cassandra more appealing to those organizations, Cassandra supports a query language that is semantically similar to traditional SQL.
While it’s unlikely that Cassandra will replace Oracle or IBM relational databases any time soon, there has been a noticeable decline in new database licenses for both vendors. The degree to which Cassandra may be responsible for that is difficult to assess.
Cassandra as a service can also be found on Amazon Web Services and a variety of other hosting services, including one managed by DataStax itself. Earlier this year DataStax also partnered with HP to provide HP Moonshot servers that come prepackaged with DSE.
In addition to the Microsoft and Rackspace announcements, DataStax today also announced the general availability of version 4.8 of its Cassandra implementation which among other things adds tighter integration with Apache Spark, the popular open source in-memory computing framework, and Docker containers. DataStax also announced Titan 1.0, a graph database that is based on a distributed framework that like Cassandra scales out more easily across x86 servers.
Robin Schumacher, vice president of products for DataStax, said another big advance in DSE 4.8 is that Cassandra now consumes 50 percent less storage space.
In both cloud and hosting scenarios, IT organizations that don’t have a lot expertise when it comes to deploying Cassandra are being offered the option of relying on third-party expertise to run both Cassandra and the applications that run on it. In fact, in an environment such as Azure, Shumacher said, DSE is the only true distributed database available, because Microsoft SQL Server is based on a master-slave architecture.
Corrected: A previous version incorrectly said Robin Schumacher was VP of product at Docker. He is VP of product at DataStax. DCK regrets the error.