As OpenStack has matured over the past 5+ years, we've grown a healthy ecosystem of ways to consume OpenStack: there are distributions, appliances, managed offerings, public clouds, and consulting services to help you build your cloud. OpenStack can be used for private clouds, public clouds, and hybrid cloud scenarios, not to mention use case specific purposes like NFV. As the software and demand for it have matured though, interoperability between OpenStack clouds has become increasingly imporant. The DefCore Committee has now released several Guidelines that serve as interoperability standards, and the Technical Committee and Board of Directors are considering changes to OpenStack's mission statement that highlight interoperability as a core value.
In this talk, we'll look at the top barriers to interoperability OpenStack faces today, and discuss what's being done to address them. We'll also review the latest work from the DefCore Committee and look ahead to it's next challenges.
First, we'll talk about what the concerns are: what does "interoperability" mean? Does it mean API's are consistently supported? That toolkits and SDK's work across many clouds today? That image formats are consistently supported? That roles and policies are universal? That workloads can be moved from one cloud to another unchanged? That OpenStack workloads can be moved to other kinds of clouds? We'll provide information from community members, Board and TC members, and the DefCore Committee on different facets of interoperability.
Next, we'll discuss what the major barriers to interoperability today are. These will include notes from some of the research the DefCore Committee is currently working on to assemble a report for the Board, Technical Committee, and User Committee.
Finally, we'll discuss how the community is addressing these concerns today, and how interested parties can help. We'll point to examples of work done by project teams, the TC, DefCore, and more.