The OpenStack Foundation (OSF) used its recent Open Infrastructure Show (OIS) to remind the open source community of its importance, maturity, and flexibility. But the event also showed that the group understands that the virtualized infrastructure environment is evolving rapidly.
I must admit that heading into the OIS event I was not expecting much. Conversations I have had over the past year continued to show a strong core of OpenStack supporters, but it seemed that the platform’s innovative spirit was diminishing. And in such a rapidly evolving technology segment, any sort of diminishing momentum is the equivalent of going backwards.
However, the OpenStack Foundation hit the ground running with a strong push around new projects targeted at growing segments of the virtualization ecosystem. This included a lot of talk on the Kata container project and the Zuul continuous integration/continuous delivery (CI)/CD) platform.
It then kept on running by embracing a more open approach to working with other organizations in the open source community.
Jonathan Bryce, executive director of the OpenStack Foundation, used his opening keynote to tout the organization’s more collaborative tone. “The only way we create a world running on open infrastructure is together,” Bryce said during a presentation with OSF Board Chairman Alan Clark. “Our community is big, and the problems we’re solving [in open infrastructure] are significant.”
Big words that showed OSF was aware of its surroundings.
One of the examples they gave was work that OSF has been doing with the Kubernetes ecosystem. Speakers noted that 61% of OpenStack users were running Kubernetes on their OpenStack deployments, and that the OSF had increased its focus on making sure that those two platforms would play nice with each other.
Now, the OSF can probably lay claim that the push to embrace the broader ecosystem as being something that was internally driven and based on OpenStack hitting a certain maturity level. But I also have to think that the group had seen a number of those other organizations gain significant traction over the past year and that it needed to embrace those efforts or it might get lost in their wake.
It’s not that OpenStack had ever lost its relevancy to telecom operators or service providers. In fact, more of those customers have probably signed up for OpenStack-based services over the past few years as they have been migrating their networks toward greater software control.
But, in talking with some of those operators and even some of their vendors, it’s apparent that many would likely either minimize or completely bypass that OpenStack approach for a more cloud-native containerized deployment architecture if they could.
“It’s almost like OpenStack is boring and everyone is looking to Kubernetes now,” said Sandro Mazziotta, director of NFV product management at Red Hat, in an interview during the show. “The discussion has really pivoted.”
OpenStack and Kubernetes
The most obvious alternative is the burgeoning Kubernetes ecosystem, which has forced just about every cloud-related company to replace the letter “C” with “K” in all of their acronyms. (That’s how you know it’s powerful.)
However, that desire continues to be hampered by a lack of maturity. This is especially important for running parts of a telecommunications network in production.
Canonical CEO Mark Shuttleworth, for one, explained that it would be “naive” at this point to expect a telecom operator or a bank to use Kubernetes alone to power their network in a production environment.
This angle was echoed by Amy Wheelus, vice president of AT&T’s Network Cloud, who also cited the need for the cloud native space to gain more maturity before it could be considered an all-powerful option. The carrier is helping those efforts by using both Kubernetes and OpenStack as part of its Airship deployment. Wheelus made it clear that, “OpenStack is critical to our success and what we are doing with Airship.”
Mazziotta said that he expects telecom operators that are virtualizing parts of their infrastructure to support 5G services to initially rely on OpenStack and more traditional virtual machines (VMs). However, he added that he then expects Kubernetes and containers to be a growing part of those deployments during the second half of next year.
“It’s really a safety net to do a trial on OpenStack and then move to production,” Mazziotta said. “Then I see them moving to Kubernetes and containers.”
The Future Is Coming Fast
The cloud native community is also racing to speed up that adoption curve.
For instance, the Linux Foundation and the Cloud Native Computing Foundation (CNCF) recently launched a test bed platform that is designed to speed the rise of Kubernetes as a test platform. The test bed’s purpose is to show the ability to run the same networking code running as virtual network functions (VNFs) on OpenStack and as cloud native functions (CNFs) on Kubernetes.
For telecom-focused vendors and operators, the test bed allows for a consistent and reliable method to test network functions from ONAP or an organization’s own networking code as VNFs and CNFs. Those organizations can then compare the performance and resiliency between running on Kubernetes and OpenStack using the same underlying hardware.
A related chart released by the Linux Foundation and CNCF illustrates that initiatives real motive:
Notice that as time evolves, Kubernetes squeezes OpenStack out of the picture? Sure, this is coming from a group that will benefit from such a squeeze, but it also aligned with what many at the event were hinting at.
I think the OpenStack Foundation did a great job at the OIS event in re-establishing its position in the ecosystem hierarchy. And it’s likely that the moves to expand its reach will pay dividends. But, it will be key for the group to continue pushing innovation into its platforms because it’s obvious there are others out there looking to take advantage of any hiccups.