OpenStack allows businesses to manage their own private clouds, like AWS, on-premises. After 29 releases, it remains one of the world's most active open source projects, and this week the OpenInfra Foundation, which leads the project, announced the release of OpenStack version 29. The new release, dubbed “Caracal,” highlights new capabilities for hosting AI and high performance computing (HPC) workloads.
The typical OpenStack user is a large enterprise. It could be a retailer like Walmart or a major telecommunications company like NTT. What almost every company has in common today is considering how to bring AI models into production while keeping their data safe. For many, that means maintaining full control of the entire stack.
As Nvidia CEO Jensen Huang recently pointed out, we are in the midst of a multitrillion-dollar wave of investment in data center infrastructure. Most of that investment is from the big hyperscalers, but a lot of it is also going into commercial deployments, and those data centers need a layer of software to manage them.
As such, OpenStack is currently in an interesting position as one of the only comprehensive alternatives to VMware's offerings, but with many VMware users not too happy about the sale to Broadcom, OpenStack has its own problems. facing. Now more than ever, VMware users are looking for alternatives. “With Broadcom's acquisition of VMware and some of the licensing changes they've made, a lot of companies are coming to us to reconsider OpenStack,” said Jonathan Bryce, executive director of the OpenInfra Foundation. “He came to visit us,” he explained.
Much of OpenStack's growth in recent years has been driven by adoption in the Asia-Pacific region. In fact, as the OpenInfra Foundation announced this week, its newest Platinum member is Okestro, a Korean cloud provider with an emphasis on AI. But Europe, with its strong data sovereignty laws, is also a growing market, such as the UK's Dawn AI supercomputer running his OpenStack.
Mark Collier, COO of the OpenInfra Foundation, told TechCrunch that “everything is in place for a major turnaround to infrastructure and the adoption of open source.” “This primarily means OpenStack, but also includes Kata Containers and other projects. Therefore, we expect a new wave of infrastructure upgrades to be completed in our community over the coming years. It’s very exciting to see them being given important work to do.”
In fact, new features added in this release include support for vGPU live migration in Nova, OpenStack's core compute service. This means that a user can move his GPU workload from one physical server to another with minimal impact on the workload. This was a request from businesses who wanted to be able to manage their expensive GPU hardware as efficiently as possible. CPU live migration has long been a standard feature of Nova, but this is the first time it's also available for GPUs.
The latest release also includes a number of security enhancements, including rules-based access control for more core OpenStack services like the Ironic bare-metal-as-a-service project. This comes in addition to network updates and numerous other updates to improve support for HPC workloads. Complete release notes can be found here.
This update is also the first since OpenStack moved to the Skip Level Upgrade Release Process (SLURP) a year ago. The OpenStack project releases new releases every six months, which is too fast for most enterprises. Early in the project, most users described the upgrade process as “painful” (or worse).
Upgrading is now much easier and projects are much more stable. SLURP's pace introduces something similar to a long-term release version. Every year, every second release becomes an easy-to-upgrade SLURP release, even though the team still makes major updates in the original 6-month cycle. People looking for a faster cadence.
Over the years, OpenStack has gone through cycles of ups and downs in terms of recognition. However, it is now a mature system, supported by a sustainable ecosystem. That wasn't necessarily the case at the height of the first hype cycle a decade ago. In recent years, the company has achieved a lot of success in the telecommunications industry, allowing it to pass through this stage of maturity, and it now appears to be in the perfect place and timing to take advantage of the AI boom.