OpenStack sets its sights on the next generation of private clouds

Today, the OpenStack Foundation is launching the latest version of its platform that allows enterprises to run an AWS-like cloud computing platform in their data centers. Ocata, as the 15th release of OpenStack is called, arrives after only a four-month release cycle, which is a bit faster than its usual six-month cycle, which will resume after this release. The reason for this is a change in how the Foundation organizes its developer events over the course of a release cycle. Because of the shorter cycle, this new release focused more on stability than new features, but it still manages to squeeze a number of new features in as well.

At this point, OpenStack is a massive project that consists of almost 20 sub-projects. All of these are getting updates, of course, but what stands out here is that many of these new feature focus on better support for software containers in OpenStack. As OpenStack COO Mark Collier told me, the container projects are growing faster than the other projects. He described the combination of OpenStack with the Google-incubated Kubernetes container orchestration system as the “LAMP stack of the cloud” and attributed Kubernetes’ popularity to the fact that Google was willing to relinquish control over it and allow an open ecosystem to grow around it that wasn’t dominated by a single company.

For the Octa release, improved container support means better integration of Kubernetes into Kolla, the Openstack project that aims to make OpenStack itself easier to deploy on containers. The advantage here is that this doesn’t just make managing an OpenStack deployment easier, but also makes upgrading it a less complicated affair. Other updates include better support for Mesosphere in Magnum, OpenStack’s main project for making container orchestration services part of its stack, as well as Docker Swarm support for its Kuryr container networking service. OpenStack is clearly not playing favorites when it comes to container engines.

OpenStack is clearly not playing favorites when it comes to container engines. Even a year ago, there was a lot of discussion about whether containers might actually spell doom for OpenStack. For the most part, those fears were likely a bit overblown and containers are now an integral part of the project.

2017-02-21_1746

In discussing the future of OpenStack, Collier also noted that he is seeing a major shift in how enterprises are looking at their private clouds. The first generation of private cloud services, including OpenStack, weren’t very easy to us. “It required a much bigger team and you saw the adoption fit that pattern,” he said. “You saw PayPal and Walmart adopt it. It was great if you were one of those companies but it wasn’t worth the time for the mere mortals.” Now, as we’ve hit what Collier considers the second generation of private clouds, that’s less of an issue and you don’t need massive teams to set up a private cloud anymore — and there’s now a robust ecosystem of companies that can help you set it up, too.

In the earlier scenario, the amount of manpower needed to set up an OpenStack cloud made it difficult to do so for small teams, but Collier believes that what we’re seeing now is an inflection point where private clouds can once again compete on price with large public cloud services like AWS. There, you tend to pay a premium for flexibility, but using the likes of OpenStack can be cost-effective for sustained workloads now.