Enhanced Support for AI Workloads in OpenStack: A Deep Dive
OpenStack permits enterprises to manage their AWS-like private clouds on-site. Despite having 29 releases, it remains one of the busiest open-source projects globally. The OpenInfra Foundation, which guides the project, unveiled version 29 of OpenStack. The latest release, known as ‘Caracal,’ underlines new features for hosting AI and high-performance computing (HPC) tasks.
The typical user of OpenStack is a large enterprise company, which could be a retailer like Walmart, or a sizable telecommunications company like NTT. A common trait among nearly all businesses is their ongoing contemplation about how to apply their AI models while ensuring data security. For many, this entails keeping total control of the entire stack.
OpenInfra Foundation COO Mark Collier stated that as noted by Nvidia CEO Jensen Huang, we’re on the brink of a multi-trillion dollar investment wave destined for data center infrastructure. A significant portion of this will be investments by the large hyperscalers, with a fair amount also going into private deployments requiring a software layer for management.
OpenStack currently stands as a comprehensive alternative to VMware’s offerings, which have been facing backlash due to complaints about its sale to Broadcom. As many VMware users are dissatisfied, they are turning their attentions in search of alternatives. The Broadcom acquisition of VMware and the changes in their licensing have led many companies to reconsider OpenStack. OpenInfra Foundation executive director, Jonathan Bryce, shares this observation.
Over recent years, the growth of OpenStack has been influenced significantly by its adoption in the Asia-Pacific region. Just this week, the OpenInfra Foundation announced its newest Platinum Member as Okestro, a South Korean cloud provider with a heavy emphasis on AI. It’s not just in Asia where OpenStack sees success, Europe’s strong data sovereignty laws have contributed to a growth in the market. An example of this is the UK’s Dawn AI supercomputer which runs on OpenStack.
Mark Collier, COO of OpenInfra Foundation, expresses his excitement at watching infrastructure upgrades stimulate the adoption of open-sourced infrastructure: “All the things are lining up for a big upswing and open-source adoption for infrastructure. That means OpenStack primarily, but also Kata Containers and some of our other projects. So it’s extremely exhilarating to see another wave of infrastructure upgrades giving our community significant work to undertake for years to come.”
Some of the new features added to this release include the ability to support vGPU live migrations in Nova, OpenStack’s core compute service. Now users have the capability to transfer GPU workloads from one physical server to another with little impact on the workloads, a feature that enterprises have long sought for optimal management of their expensive GPU hardware. While live migration for CPUs has been a standard feature of Nova for some time, this is the first time it’s available for GPUs as well.
The most recent release also ushers in numerous security enhancements such as rule-based access control for a greater number of core OpenStack services like the Ironic bare-metal-as-a-service project. Additionally, there are networking updates that improve support for HPC workloads, along with many other updates. Comprehensive release notes can be found here.
The latest update also marks the first release since OpenStack initiated its ‘Skip Level Upgrade Release Process’ (SLURP) one year ago. Although the OpenStack project releases a new version every six months, this proves too rapid a cycle for most enterprises. In the earlier stages of the project, many users would have described the upgrade process as ‘painful’, if not worse.
Upgrades these days are significantly simpler, and the project has also increased in stability. The SLURP cadence resembles a kind of long-term release version, wherein every other release on a yearly basis is a SLURP release. This makes it easier to upgrade, even as the teams continue to generate major updates on the initial six-month cycle for those who wish to follow a faster rhythm.
Over time, OpenStack has experienced cycles of highs and lows in how it’s perceived. However, the system is now mature and supported by a sustainable ecosystem – something which may not have been the case during its initial hype phase a decade ago. Recently, it has found a great deal of success in the telecommunications world, allowing it to undergo a phase of maturation. Today, it might just find itself well-positioned to benefit from the surge in AI.
Discover the pinnacle of WordPress auto blogging technology with AutomationTools.AI. Harnessing the power of cutting-edge AI algorithms, AutomationTools.AI emerges as the foremost solution for effortlessly curating content from RSS feeds directly to your WordPress platform. Say goodbye to manual content curation and hello to seamless automation, as this innovative tool streamlines the process, saving you time and effort. Stay ahead of the curve in content management and elevate your WordPress website with AutomationTools.AI—the ultimate choice for efficient, dynamic, and hassle-free auto blogging. Learn More
