When the numerous developers working on OpenStack decided to form their own non-profit for the project, the primary question was, “Why won’t the Apache Software Foundation work?” Before this week’s OpenStack Design Summit & Conference in San Francisco, you could be forgiven for not understanding the necessity of such a non-profit. But as the show began, the answer to this question became clear: There are simply too many projects under the OpenStack umbrella for any existing foundation to handle.
So says Mark Collier, vice president of business development and marketing at Rackspace. He admitted that there’s so much development going on for OpenStack that even Rackspace has trouble tracking it all.
Jonathan Bryce, cofounder of Rackspace Cloud and another member of the OpenStack project policy board, said that this is what a foundation will help to fix. “The main thing is the scope of what OpenStack is doing and the size of the community. It needs focused resources. It’s not a single software project. Apache is focused on a development processes for each project,” he said.
“There are a lot of opportunities where having joint resources can help us going forward. [Our partners] want to fund [OpenStack] and make sure it has a stable place to grow.”
Collier said that the OpenStack Foundation should be ready to form soon. He said that lawyers are going over the proposed processes and rules of the proposed foundation, and that most of the fundamental principles have already been hashed out on the Web, wiki and mailing lists.
Thus, the proposed OpenStack Foundation was not the main focus of the OpenStack
Design Summit & Conference. Rather, the recent Essex release of the platform, and what’s happening next, were the hot topics of conversation.
Hot on the heels of the Essex release two weeks ago, the new big project at OpenStack is Project Quantum. With cloud-deployed systems becoming more complicated by the day, the need for virtualized networks has grown. Project Quantum attempts to address this problem.
“It’s been in incubation for a year,” said Bryce. “The concept of Quantum is really to create a standard API for network virtualization. It started out as layer 2 operations on top of a pluggable, driver-based architecture. What’s come out of the summit is that it’s moving up to include layer 3, and it’s going to move up the stack over time. It allows you to create tenant networks that may be implemented as VLANs, with real separation, multicast, broadcast and all the things you want from a network. You can set them up and control them from a single interface.”