Over the past 9 months or so, I’ve written two Standards Watch blogs on the topic of OpenStack: A Big Year for OpenStack; and NetApp and OpenStack, which I posted around the time of the OpenStack Summit in October 2012. Since then I have the impression that the initiative has advanced dramatically. The code releases are continuing on their regular 6-month cadence; the hype hasn’t abated in fact it seems to have gained momentum; almost all customer conversations about internal cloud infrastructure include the topic of OpenStack; and now we are hearing more about production deployments. Consequently I’ve have been looking forward to this week’s OpenStack Summit in Portland.
Beth Pariseau wrote a good pre-show piece in SearchCloudComputing focused mostly on the new features in the latest release of OpenStack (named Grizzly), but I wanted to assess the whole thing for myself. In particular I was looking a sense of the commitment of the OpenStack Foundation member companies, the state and quality of the latest code release (Grizzly), customer engagement and the velocity of deployment growth.
OpenStack Foundation Member Commitment
At the last Summit, the OpenStack Foundation has only just recently been launched, but much of the focus of the event was on the development community, as it had been for previous events. Apart from the fact that Platinum and Gold Members had shown their commitment to the Foundation by paying the hefty dues, it was tough to figure out what their adoption plans were.
This week’s event was quite different. Most of the Platinum and several of the Gold member companies made extensive presentations about their commitment, productization and/or deployment plans for OpenStack. In addition a wide range of member companies talked about their contribution plans in the panel sessions.
With almost 4000 people in attendance, most of them developers, this commitment also made this the biggest OpenStack Summit yet. There’s no question in my mind that a broad range of companies are committed to making OpenStack the most ubiquitous cloud platform in the industry. I didn’t have that sense before.
The Grizzly Release
At the last Summit, much of the buzz was on the much-anticipated new features of the latest release (Folsom). Some of the reason was that previously there had been big gaps in functionality that were a barrier to deployment. For the latest release, there is interesting new functionality, but most of it is not filling critical gaps, but bolstering scalability and availability, and broadening the range of OpenStack-enabled products – features that may not matter so much in pilot deployments, but do matter for production deployment. Interestingly, though, there was an equal amount of buzz on code quality and resiliency. The OpenStack Foundation has dramatically increased the focus on QA in the development process, with numbers like a 600% increase in code test coverage being mentioned. The community is getting serious about releasing production-deployment quality code.
This is where this week’s Summit was dramatically different from the previous one. At an analyst panel during the October event, a recurrent question was “where are the deployments”. The same panel this year was saying “OK – we see the deployments, what workloads are missing”.
The OpenStack overview said there are now 100 production deployments of OpenStack around the world. The conference agenda featured case studies from (in no particular order) HubSpot, Samsung, Best Buy, PayPal, RackSpace, MercardoLibre, NSA, CERN, Bloomberg, Comcast, Enter.it, KIO Networks, IBS Datafort, NTT and EIG. Many of these are production deployments; some are on the cusp between pilot and production. I believe that this represents a critical mass that will accelerate customer interest in OpenStack.
However, deployment isn’t for the faint of heart. It’s complicated. You’d be advised to have an excellent team well-versed in open source, good partners, engagement with the OpenStack development community, and a phased deployment strategy. But I expect to see a dramatic increase in deployments in the coming 12 months.
Bottom line on all of this … I see more and more evidence that OpenStack could become the most ubiquitous means of building open cloud environments. It’s still early days – 100 production deployments doesn’t make it mainstream – but the pace of maturity and acceptance over the past six months has been truly impressive.