Basho Taps into Open Source Mesos Cluster Manager

Peel back the data center covers inside many of the better-known web-scale companies such as Twitter, Airbnb and Apple, and the one thing they have in common is heavy reliance of open source cluster management software known as Mesos. Originally developed at the University of California at Berkeley, Mesos makes it easier to share infrastructure resource in way that can be managed granularly.

Meanwhile, at the other end of the IT spectrum, there’s been a rise in interest of high-end NoSQL distributed database as an alternative to distributed relational database. Case in point is the Riak KV distributed database developed by Basho Technologies. Now Basho Technologies, with help from Cisco Systems, is moving to bring Riak KV and Mesos together in a way that provides a foundation for the next generation of distributed computing applications.

Basho Technologies CEO Adam Wray said that while adoption of Mesos and Riak KV is still relatively slight, it’s pretty clear that the combination of the distributed cluster management software and distributed database software is going be critical in terms of enabling the development of truly heterogeneous hybrid cloud computing. For all the talk these days about hybrid cloud computing, the management of those clouds is for all intents and purposes semi-autonomous. That's because there is no real common layer of distributed computing infrastructure that unifies resources across multiple private and public clouds.

Naturally, the combination of Mesos and Riak KV may not be the only way to drive federated hybrid cloud computing. But Wray said together the two technologies clearly complement one another in terms of the scale and scope need to deliver distributed data services at web scale. In fact, Wray noted Cisco is helping drive this integration to help drive further adoption of its hybrid Intercloud initiative.

For solution providers the emergence of that layer of distributed data service will be critical to building of next-generation IT solutions. Without it, they will be left to their devices in terms of trying to stitch together a distributed data service on their own. Given the complexity of that challenge, the number of truly web-scale distributed applications that are likely to get deployed is fairly limited.

Obviously, not every solution provider has the skills needed to master Mesos and distributed databases such as Riak KV. But as solution providers look over the horizon a bit to see how enterprise computing is about to evolve, chances are high that a distributed data service resembling Mesos and Riak KV are going to a lot more commonplace than they are today.