OpenStack Juno Cloud Features Trove Database-as-a-Service Updates
When the open-source OpenStack Icehouse platform was released in April, the Trove database-as-a-service project was one of its key new features. Fast forward six months, and the OpenStack Juno release is set to debut on Oct. 16, complete with a long list of updates and improvements to Trove.
The Trove project first began in 2012 under the name Project Red Dwarf—a joint project led by Hewlett-Packard and Rackspace. Since then, database virtualization vendor Tesora has become actively involved in Project Trove and is now one of the leading contributors to the project. Tesora also offers community and commercially supported versions of Trove to users.
Ken Rugg, CEO of Tesora, told eWEEK that when he first looked at Project Red Dwarf, it was clear to him that the project still had a long way to go. With the inclusion of the renamed Project Red Dwarf as Trove in the OpenStack Icehouse release earlier this year, the project has accelerated.
With the upcoming Juno release, both new features and new database support have landed, Rugg said. In the Icehouse release, support was primarily focused around the MySQL database. In the Juno release, Trove now supports the open-source PostgreSQL database and includes better support for multiple NoSQL databases, including Apache Cassandra, MongoDB, Couchbase and Redis.
From a features perspective, Trove in the OpenStack Juno release now includes data replication support for MySQL.
"The replication in this release is actually read replicas," Rugg explained. "That's the ability to spin up multiple read copies from a master database."
Currently as a project, Trove does not actually get into the database's data path, Rugg added. From a database replication perspective, what that means is that automated failover for high availability is not something that Trove can currently enable.
"So Trove itself doesn't get into the data loop, where it can be operating as a proxy to enable transparent database failover," he said.
That said, the database failover can be enabled by way of a third-party orchestration tool or even with the OpenStack Heat project for cloud orchestration.
Trove in OpenStack Juno will now also support database clustering for the open-source MongoDB NoSQL database. Sharding is a database technique whereby a large database can be divided up, or "sharded," into multiple smaller elements to provide better database performance.
Support for the Neutron networking project within OpenStack is also new to Trove in the Juno release cycle. Neutron provides a software-defined network (SDN) as a service layer to an OpenStack cloud deployment.
"With the new Neutron support, an administrator can specify what virtual NIC [Network Interface Card] a database should be using," Rugg said.
Moving forward beyond the Juno release, there are already some ideas as to where Trove should head for future OpenStack releases. Rugg said there is the future potential to support Oracle Database as well as Microsoft's SQL Server. Currently, not all of the databases have the same level of feature support within Trove, which is another area for future improvement.
"We need to get to the point where all that databases have all the different features, including backup, replication and configuration options," Rugg said. "We need to continue to build Trove out and make it more industrial-strength."
Sean Michael Kerner is a senior editor at eWEEK and InternetNews.com. Follow him on Twitter @TechJournalist.
Read more on "Server Virtualization Spotlight" »