Six months on from its announcement at Openstack Summit Sydney in late 2017, community testing project Openlab is in full swing.
Openlab was initially formed by Intel, Huawei and the OpenStack foundation as a community-led project for improving SDK support and also introducing other platforms like Kubernetes and Cloud Foundry to the Openstack environment. Ultimately the idea is to improve usability in hybrid and multi-cloud environments.
Melvin Hillsman sits on the governance board along with Dr Yih Leong Sun of Intel and Chris Hoge from the Foundation. Hillsman moved from Rackspace to Huawei to work specifically on the project.
"The reason we think Openlab is important is, basically, Openstack for some time has been very specific about testing and integration for Openstack services, focusing only on the projects started at Openstack," Hillsman tellsComputerworld UK at the Openstack Vancouver Summit. "It's been working very well, it's a robust system. But for me as a person in the user community - my getting involved in Openstack was more on the operator-user side.
"So it would be great to have some of these deployment integrations, features, ideas, thoughts and communications for users - the community above where the Openstack APIs end."
Above the SDK layer there are then other industry verticles, for example Kubernetes, which Openlab had worked on to get up and running with Openstack and other architectures at once.
Hillsman describes the effort as a group of people in that "broader ecosystem coming together and saying: 'OK, we have a use case, we are willing to collaborate, we know the technology, and now we need somewhere to test this stuff'."
"That's kind of the idea - to ensure that when I want to use a number of different things to get a service, I don't have to worry about if it's been tested well or if it works well. If I deploy it, and something's broken, our hope is it's not any of the projects we are trying to use."
The other driving point is one that's been echoed in the Openstack community - that it's not a handful of companies in control, or if it is technology that's being developed behind closed doors, Openlab tries to get it into and developed by the community.
"It matters because users matter," Hillsman says. "If the software is great but no one is using it, what's the point? You have to innovate and ensure that the users concerns, their experiences, are communicated and validated. It's focusing on the users."
And what would success look like for the fledgling project?
According to Hillsman, Openlab has grown "quite a bit" since it was announced last November. But success means getting more companies and more users involved, including the participation of product managers and cloud architects. The group demonstrated a Kubernetes-Openstack integration pushed out onto an ARM server during the opening keynote, and they hope to be able to show off another integration in six months' time at the next Summit, in Berlin, Germany.
"Further term, it's for someone else to step in and give other folks an opportunity to lead the effort," says Hillsman. "Under the context of user voice, and validation, and integration, and to give them a chance to help shape what Openlab means and the different architectures that should be added.
"We'd love to have some Power from IBM, we'd love to get into some IoT things. Just getting as much information as we can for some solid integration we can shop off, that's a cycle, wash, rinse and repeat - these integrations being shown and available for others to try and use and take further, and then we go to the next one. It's bigger than just a few people being able to handle it."
That chimes with the essential message of open source, that transparent collective efforts towards achieving a common goal can deliver development in some areas unprecedented compared to isolated organisations working away at the same task.
Nevertheless, it's a "balancing act", says Hillsman: "You have to be careful because, of course, commercial entities have commercial concerns. And it's OK, I think the openness of these entities to say: 'Hey, we realise that some layers - like the internet - we can't corner. We can't corner the internet, we can't corner infrastructure as a service, we can't corner these different things.
"However, if we work together where the value prop is not that big... we can then start building services on top, and differentiate."