Pivotal Software this week announced it has made the Greenplum massively parallel processing (MPP) database available in beta on Kubernetes.
Announced at the Greenplum Summit, the forthcoming version 6 of Greenplum will take advantage of multiple pods with a Kubernetes cluster to process database queries in parallel.
Jacque Istok, vice president of data at Pivotal, says Greenplum 6 will be available on multiple distributions of Kubernetes, including the Pivotal Container Service (PKS) instance of Kubernetes that Pivotal developed with sister company VMware. DevOps teams will be able to take advantage of the BOSH framework originally developed to automate the deployment of the open source Cloud Foundry platform-as-a-service (PaaS) environment to deploy instances of Greenplum and Kubernetes, he says, noting the combination of BOSH and Kubernetes will have a critical role in also making it easier for database administrators to keep pace with rapidly changing requirements in the age of agile development and DevOps.
At the same time, Pivotal continues to work on rearchitecting Greenplum to run on top of an open source instance of the Postgres relational database. Rather than reinventing every element of a distributed relational database from the ground up, the latest generally available version of Greenplum makes use of Postgres as the foundational element of the MPP database. Pivotal will then continue to focus its efforts on add MPP services on top of Postgres. The current Greenplum 5 offering is based on version 9.4 of Postgres.
Pivotal this week also announced it will make available a commercially supported instance of Postgres alongside the current Greenplum 5 platform. Greenplum itself will be updated to stay current with more recent releases of Postgres, says Istok. Pivotal already makes available a commercially supported instance of the open source MySQL database, but since Oracle acquired MySQL, many IT organizations have been transitioning to Postgres.
As Pivotal moves to essentially containerize the engines that make up its MPP database using Postgres, there’s a significant opportunity to expand adoption of MPP databases. Most databases today don’t take full advantage of modern multi-core processor platforms. MPP databases increase utilization of those platforms by distributing processing engines across the platform. Istok says customers have been asking Pivotal to make Greenplum available on top of Kubernetes to make it easier for them to take advantage of a platform through which it becomes easier to manage MPP database engines running in multiple Kubernetes pods.
Support for Kubernetes is part of a long-term Pivotal effort to unify transactional and operational analytics workloads. While these workloads today are still for the most part processed in isolation, more organizations now want to be able to process analytics in real-time alongside transactions to make more informed decisions in real time. Traditionally, however, the underlying infrastructure did not make it simple to process different types of workloads in parallel unless there was a complex proprietary MPP database in place, which tended to be very expensive to acquire, deploy and manage.
Kubernetes and BOSH promise to reduce that complexity, while Postgres should help drive down overall costs. Naturally, it’s too early to say whether there will be a resurgence of interest in MPP databases, but all elements required to spur increased adoption are starting to fall into place.