Telecom network operators have a love-hate relationship with OpenStack. On the one hand, it is ubiquitously deployed as an open-source cloud platform, and the default choice for network functions virtualization infrastructure (NFV-I). Some operators have moved forward aggressively, even bragging about the breadth of their OpenStack deployments. Others have been as open in complaining about what they saw as the open source software's challenges around scalability, security, and backward compatibility. (See OpenStack Underpins Comcast Elastic Cloud, AT&T Rallies Carriers Around OpenStack, and BT Says OpenStack Still Not Up to Spec.)
With the growing emphasis on edge computing, however, telecom is once again pushing OpenStack forward as a likely solution, but this time the OpenStack Foundation is taking a fundamentally different approach. According to Executive Director Jonathan Bryce, there is a very different process now at work within his organization to tailor OpenStack work to specific use cases, one of which is specifically focused on edge computing.
This change in approach began in 2017 with a strategic assessment of where and how OpenStack was being used, including a growing push of the software into locations well outside of the traditional data center environment, Bryce tells Light Reading in an interview. That assessment came from the realization that OpenStack was being used in ways that went well beyond its usual scope.
"The community that had formed up around OpenStack had moved beyond the technology to containers and NFV, and even to some extent, serverless frameworks and all of these kinds of workloads were running inside the OpenStack environment," he says. "So what we did last year was to kind of reorient the OpenStack Foundation around use cases that are based on the core compute, storage and networking principles but that take into account what those differences are."
The first strategic focus areas are the data center/cloud -- the OG OpenStack -- container infrastructure, edge computing and NFV, and CI/CD or continuous integration, continuous development, Bryce says. Containers were a no-brainer given that about half of the OpenStack clouds already run Kubernetes for container management, he adds. The NFVi space was another obvious choice, while edge computing was what telecom operators -- among others -- were saying they wanted most.
"We didn't choose edge computing, it chose us," he comments.
All of these use cases have basic needs for automated compute, storage and networking, but there are variations in how those things are accomplished. Also, there are more specific requirements for some use cases.
Edge computing, for example, needs self-service capabilities, automation and application programming interfaces, along with basic elasticity. But since edge compute will largely live outside the data center in some pretty unfriendly places -- like a cell site or street cabinet -- it has other requirements.
"It might not be in a place that human techs can go, so that introduces some constraints around the reliability and security that you need to have in there," Bryce notes. "You need to have certain isolation, recovery modes that don't require humans. And then you also have constraints on capacity and resource availability. You aren't talking about hundreds of servers; you are talking about tens of thousands of servers. Telecom companies may argue about who has the biggest OpenStack deployment, but it also matters who has the smallest."
The OSF Edge Computing Group issued its first release in February and there is ongoing trial work involving AT&T, SK Telecom and others, that is moving forward. (See Open Source Zeroes In on the Edge.)
"One of the big issues at the edge is that you don't have one environment, you have a lot of different environments, so that automation, that zero-touch management is very critical so this is being built directly off of AT&T and SK Telecom requirements and both of them are trialing it," Bryce says.
The OpenStack Foundation will be looking closely at the new tools and requirements for this "massive multi-cloud" environment, he adds.
"The community aspect is critical, the working group is really critical because it keeps the feedback loop going," Bryce says.
- Open Source Zeroes In on the Edge
- OpenStack & NFV Help Drive Red Hat Growth
- ONF Operators Take Charge of Edge SDN
- OpenStack 'Queens' Takes on New Workloads
- AT&T Sharpens Edge With New Open Source Effort, Test Lab Launch
- Operators Must Cloudify at the Edge
- Red Hat OpenStack Platform 12 Eyes Greater Container, Security Support
- How OpenStack Enables the Telco Cloud
— Carol Wilson, Editor-at-Large, Light Reading