While Red Hat understandably wouldn't discuss anything to do with the IBM mega-acquisition - for regulatory reasons - director of product management James Labocki and senior director for product management at Openstack Nick Barcet confirmed Red Hat's commitment to Openstack for at least the next "10 years".
The open source giant today announced Openstack Platform 14, which Barcet said aimed to make Openstack a better platform to run container orchestration system Kubernetes on, while also helping to better manage the deployment of Red Hat's container platform OpenShift on bare metal, as well as easing the integration of OpenShift and Openstack at the networking and storage layer. This, says Barcet, is a landmark move for Red Hat because it is part of a new strategy to focus on its whole portfolio as a single entity rather than individual products.
Historically open source infrastructure project Openstack is known as a complex beast, which Barcet acknowledges."The biggest problem with Openstack is it's solving lots of problems, and different problems depending on how you configure it, which means that your initial understanding of Openstack is going to be defined by the use case you're trying to solve," says Barcet. "That can take quite a bit of time for people, and we failed as a community to make that part simpler. It would be really nice to have pre-scripted scenarios per use case you're trying to solve.
"This is something we have to improve over time."
But Barcet says that the company is seeing "lots of success" with Openstack, ranging from general purpose infrastructure, big data deployments, and calculation environments such as for AI or standard high performance computing.
"We are seeing a lot of people also using Openstack not for the 'above the cloud' aspect but the ability to deploy workloads on bare metal, and that is something that is increasing really fast," says Barcet. "More and more people are addressing their application deployment with a cloud API but directly on bare metal because they need to talk directly to some kind of hardware at the physical layer."
That might seem to go against the grain compared to many of the digital-native big brands that often shout the loudest about their infrastructure. However, Labocki adds that the company has announced new tooling to help mass migrate virtual machine-based workloads from proprietary virtualisation directly onto Openstack instead.
"Basically being able to discover them, then seamlessly move them over en masse," he says. "What we are starting to see is a lot of customers are recognising there are diminishing returns in their virtualisation infrastructure, and they're starting to say: hey, I want to invest in other technologies: Kubernetes, automation, multi-cloud management, you name it.
"We're seeing more differentiation and business value being driven out of that, as opposed to the standard virtualisation layer, which is not necessarily delivering returns any more," Labocki adds. "We're seeing a large amount of them starting to say, I want to reduce the spend here - and migrate, whether it's Openstack or another KVM-based virtualisation product."
Adds Barcet: "I don't know any Fortune 1000 company that has really migrated everything to public cloud. Some have made big announcements that they have the intention but the reality is some of their workload has to remain in-house for a variety of reasons, whether it is regulation, whether it is capacity.... It's just not happening. We are also seeing customers swinging back because the cost of op-ex from running on public cloud for a few years just doesn't make it."
When it was announced that IBM was intending to acquire Red Hat there was widely reported speculation that it was primarily to strengthen the century-old corporation with a better standing in open source and crucially, hybrid cloud.
Labocki stresses the importance of open source past, present and future.
"AI, machine learning, big data, microservices, containers, IoT, what are those built on? They're fundamentally built on Linux and open source technologies. So at the end of the day organisations, if they are going to want to digitally transform, will have to embrace Linux and open source technologies.
"The reality is that if your staff is not getting trained in that, you're going to be behind the curve on how able you are to digitally transform. And the nice thing about it is when you look at all those fundamental technologies of virtualisation and infrastructure, all those, they're built on Linux. The past is built on Linux, and the future is built on Linux, so I think it's a pretty obvious choice if you're a leader in a company you have to invest in Linux and open source skills to be able to do that."
Barcet adds that in his view Red Hat has balanced its development model and business model to achieve success, and being one of the relatively few number of highly successful open source companies.
"The development model is upstream first, it is massive investment into the community, everything is open, the business model is about delivering stability through a subscription model - and answering those complex problems that you have as soon as the software touches the hardware, because regardless of how fast we can make software evolve, hardware, the cost of racking something, hardware is always a drag. You're going to have to stabilise the platform for at least three years, maybe five years, maybe 10 years, depending on your investment model.
"Red Hat has understood that and delivered while investing at a very fast pace in the community, a subscription model that allows customers to deploy at their own pace, and maintain within the right range."
When asked about how forward looking Red Hat is able to be about the coming months and years, Barcet says that it is committed to Openstack "for at least the next 10 years" although this will probably be part of a wider 'open infrastructure' project, and stresses the importance of 5G and edge computing.
"It is going to be used for a long time. Will Openstack transform? Obviously it will but it has been transforming already - public cloud, private cloud for telcos... It's just a continued advancement. My only thing is I think we need to innovate faster nowadays than when we started the project, that's why I'm suggesting the development cycle be shortened."
"One of the big benefits of the open source model is the ability to pivot," adds Labocki. "If you go back to the Zen virtualisation-KVM virtualisation pivot that Red Hat did - no other traditional software company that had a traditional development model with proprietary development would have said we're just going to switch our virtualisation technology out in the next release. The idea that we can openly say we're going to make a move and change: you can see the same thing on the container landscape, we embraced Docker and we embraced Kubernetes, we went all in on these technologies when we already had a container technology at Red Hat.
"So when you're sitting in the audience and hear open infrastructure and you go, oh no, something is wrong - it's actually no, that's the way the open source development model works, you change quickly and you move."
What else is in the future? Barcet explains that the company is heavily involved in edge computing pilots with a number of telcos, and the ability to manage multiple data centres that are nearer the customer as well as looking into the way 5G will impact manufacturing.
"One of the big problems when you're manufacturing is you need to collect a tremendous amount of data to understand what's happening in your production chains," he says. "And sending this data over the wire may not make sense, so you also need to have a distributed model, distributed compute nodes or putting small footprint deployments even closer to your end user."