Bryce: ‘OpenStack will power the planet’s clouds’

The OpenStack Foundation announced new interoperability and testing requirements as well as enhancements to the software’s implementation of federated identity which the Foundation’s executive director Jonathan Bryce says will take the open source cloud platform one step closer to world domination.

OpenStack’s key pitch beyond being able to spin up scalable compute, storage and networking resources fairly quickly, is that OpenStack-based private clouds should be able to burst into the public cloud or some private cloud instances if need be. That kind of capability is essential if the company is going to take on companies like AWS, VMware and Microsoft, but has so far been quite basic in terms of implementation.

But for that kind of interoperability to happen you need three things: the ability to federate the identity of a cloud user so permissions and workloads can port over to whatever platforms are being deployed on (and to ensure those workloads are secure); a definition of what vendors, service providers and customers can reliably call core OpenStack, so they can all expect a standard collection of tools, services, and APIs to be found in every distribution; and, a way to test interoperability of OpenStack distributions and appliances.

To that end, the Foundation announced a new OpenStack Powered interoperability testing programme, so users can validate the interoperability of their own deployments as well as gain assurances from vendors that clouds and appliances branded as “OpenStack Powered” meet the same requirements. About 16 companies already have either certified cloud platforms or appliances available on the OpenStack Marketplace as of this week, and Bryce said there’s more to come.

The latest release of OpenStack, Kilo, also brings a number of improvements to federated identity, making it much easier to implement as well as more dynamic in terms of workload deployment, and Bryce said that over 30 companies have committed to implementing federated identity (which has been available since the Lighthouse release) by the end of this year – meaning the OpenStack cloud footprint just got a whole lot bigger.

“It has been a massive effort to come to an agreement on what we need to have in these clouds, how to test it,” Bryce said. “It’s a key step towards the goal of realising an OpenStack-powered planet.”

The challenge is, as the code gets bulkier and as groups add more services, joining all the bits and making sure they work together without one component or service breaking another becomes much more complex. That said, the move marks a significant milestone for the DefCore group, the internal committee in charge of setting base requirements by defining 1) capabilities, 2) code and 3) must-pass tests for all OpenStack products. The group have been working for well over a year on developing a standard definition of what a core OpenStack deployment is.