Days 2 and 3 of the OpenStack Design Summit & Conference kicked off on a rabid pace. These are typically the days that the deep-dive technical sessions really ramp up and it gets, well, really geeky.
The conference format was a little different this year than in the past as a new session format was introduced. Each session submitted this year had a type/distinction to which they had to adhere: "Brainstorm," "Presentation" or "Workshop." Each distinction came with time allotment windows and gave the attendees a good idea of what they would be getting into. Personally, I felt this was very beneficial, as in past years you could find a hardcore dev in a presentation when their time would be much better spent in a design brainstorm session or vice versa. Going along with usual Summit scheduling, each day there is always a sprinkle of each of the core projects along with auxiliary projects based on OpenStack, but the Quantum track appears to have ruled Day 2 based on hallway chatter and general conversations at the hotel and during the evening events.
From the Nova and Swift sides of the house it was really interesting to see the level of, dare I say, enterprise complexity in the sessions. I'll dive deeper into my thoughts about the three in future posts.
Quantum is an incubated OpenStack project that has the mission to “provide network connectivity as a service between interface devices managed by other OpenStack services (e.g., Nova).” Networking as a whole has been a pain point for OpenStack since the early days — with little flexibility in its network manager — so Quantum is being met with great anticipation. At the Boston Design Summit, the Quantum sessions were standing room only with people overflowing into the lobby. This was before there was much tangible code and Quantum was more of an idea that had a handful of developers hacking on it. Now that Quantum’s code has had more than six months to bake, and it has ramped up from a handful of devs to more than 20 with many dedicated from Nicira, these sessions were the place to be. Standing room only would be an understatement, people were fighting for spots and some attendees just didn’t even leave the room(s) all day out of fear of losing their spots.
I was particularly interested Quantum sessions due to the announcement earlier in the week in which Rackspace launched the next generation Rackspace Cloud into a Limited Availability program and unveiled its Cloud Networking product, which will be based on this exact project. In the near future, you will see Rackspace’s involvement in Quantum continue to increase as another project, codenamed Melange, is likely to be merged with Quantum for the official Folsom release. That will bring IP management and other network management services under one single umbrella. I recommend the following Quantum-based links, lots of really good stuff for operators, devs, and consumers alike:
OpenStack Compute, code-named Nova, which is under the leadership of Project Team Leader (PTL) Vish Ishaya, has grown up very quickly over the past two years. It’s really amazing to sit back and think about how far the project itself has come. There was one Twitter quote from Thierry Carrez, OpenStack Release Manager, that put it eloquently, “@tcarrez: @OpenStack community maturing: last year at design summit we had ’bzr vs. Git,’ today we have ’making configuration of OpenStack easier.’”
This of course doesn't just cover Compute itself, but it's the one I felt has come the longest way. I may be biased a bit since I deployed the Austin code for the Bexar Design Summit and that was a challenge. The overall takeaway for me is when speaking to enterprise customers now is that I can do so with more confidence today than I could even six months ago. Selling an idea to a technical audience has its easy points; they will give you the benefit of the doubt if you make sound arguments. C-level executives and vice president decision maker-types hit you hard with the business propositions - the stuff they live and breathe - and some even are just as technically savvy as their IT counterparts and love cranking on the low level stuff.
I have had more than five talks with high profile enterprise attendees and the common theme among them all was their belief that Essex, OpenStack’s fifth release, has really gotten over the “is it real, is it stable” hump and they are now taking a more serious look at OpenStack. This was strongly evidenced by an impromptu lunch announcement Wednesday during which an enterprise cloud architect spoke to the stability of Essex and shared the news Essex won a recent internal bake-off versus “another virtualization platform.” He congratulated the community for the tremendous work it has done, and admitted that it beat his own expectations for a stable release. The details of the opposing company where withheld likely for PR reasons, but he did mention they were a “large enterprise multimedia company.”
I expect that in coming months more and more companies will start to announce their intention to use OpenStack or will reach the point where they can publicly endorse it. The "early-mover" advantage will start to pay off soon enough, and there will be plenty of companies ready to talk all about it. There were a good number of sessions on Nova that really spoke to this:
I’ve only touched the tip of the iceberg in this post, as it was intended to be a primer for getting a feel for all the good talks that happened during the Folsom Design Summit. There were full tracks on topics such as Ecosystem, Horizon (OpenStack Dashboard), Keystone (Identity Service), Swift (Object Storage), Documentation, Common Development, etc… You can find a full listing of all the sessions at http://summit.openstack.org/ and the subsequent session notes or slides at http://wiki.openstack.org/FolsomSummitEtherpads.
I have to give big props to all the people that were involved in making this event a reality and a huge thank you for all the sponsors. They both truly made this, in my opinion, the most productive Summit to date, especially when it came to the community being fully engaged in not only writing excellent code but also taking on more broad business needs. I am very excited to see where this momentum takes OpenStack in the next six to 12 months. I expect we’ve only begun to see the start of the open cloud revolution.