I've been following OpenStack since its inception half a dozen years ago.
Back then the fledgling open source cloud operating system was a little project created jointly by Rackspace and NASA. A lot of time, massive amounts of funding and a few different value propositions have gone under the bridge and now OpenStack has its own foundation, huge vendor buy-in and a growing number of production deployments at scale.
Perhaps the biggest indicator that OpenStack has come of age was the (admittedly jarring) sight of Gartner analyst Donna Scott standing on stage in the first day's keynote. While Scott's talk missed the mark by a mile (she spoke at length about the Gartner bi-modal model, a hotly contested topic, but her talk was more of a conceptual piece targeted at a CIO audience, completely different to the 7,500 primarily developer-based attendees at the summit), the wiseness or otherwise of having the talk in that setting, and the fact that Gartner was prepared to go on stage and anoint OpenStack as actually a viable product, was a massive achievement for OpenStack Foundation Jonathan Bryce and his team.
This is, after all, one project which has been criticized for years for being little more than a science experiment, so an anointing by Gartner must be a satisfying end to that debate.
Luckily for the audience, the keynote also featured Boris Renski, the always colorful CMO and co-founder of pure-play OpenStack service provider, Mirantis. The thrust of Renski's talk was that while the world is all ready to suggest that Amazon Web Services, the undisputed king of the public cloud, owns the cloud generally, the revenue that AWS drives, some $10 billion, is tiny in comparison to total IT spend. The crux of Renski's point was that this is very early days and there is still a massive opportunity for other cloud vendors, and other approaches to delivering the cloud.
All in all, the keynote was interesting, and it was good to see some customers also on stage. China Mobile, Walmart, the Volkswagen Group and AT&T all got some nice stage time. Many of my analyst colleagues at the event expressed disappointment that there was a lack of technical innovation at the event generally -- I actually disagree with that perspective. I have criticized OpenStack in the past for jumping on every new technology band wagon and failing to focus on delivering a credible and stable core service first. My take from this summit was that the foundation is focusing strongly on building the core robustness of the platform rather than clutching at technology straws.
In terms of the newer customer stories on stage, Liveperson was interesting, having changed their monolithic application into 150 discreet microservices running on OpenStack with Docker and Kubernetes on large virtual machines. TCP Cloud demoed its smart city offering that has been implemented in the Czech republic. While it was primarily a story about the Internet of Things, TCP Cloud built on top of OpenStack and also used containers and Kubernetes.
Of course there were some interesting technology innovations on show. In particular Alex Polvi, the CEO of CoreOS, a somewhat scattered vendor in the broader container world, demonstrated the jarring named Stackanetes, software that packages the entire OpenStack system into individual containers that can be run on top of bare-metal while being managed with the Kubernetes system. The demo was pretty incredible and both created a far easier deployment route as well as a self-healing approach towards running OpenStack.
It did raise some questions, however, and in another post I will touch upon those issues. The core message from Polvi was that OpenStack is simply another application and hence running it on top of Kubernetes which is, after all, all about application management, makes sense.
The interesting thing for me about Stackanetes, other than the rocket-science aspects, were that it was, to an extent, an admission by the OpenStack foundation that OpenStack is simply an interim step between different modes of working -- the Foundation was seemingly admitting that cloud-native (i.e. non-virtualized) applications are a big part of the future and hence OpenStack can't be seen as the foundational layer for everything. It is unusual, and somewhat refreshing, to see people admit that there isn't, in fact, one rule to rule them all.
Around the conference floor I spoke with other customers -- Glenn Ferguson from Wells Fargo talked about the fact that for his organization, the public cloud was not an option. Rather an OpenStack, private cloud is allowing them to create agility across their legacy and greenfield applications. Wells Fargo has thousands of apps and application development teams with a variety of approaches -- from agile teams using continuous development and integration, through to more conservative teams. For Wells Fargo the important point was to embrace a platform that would work with whatever comes next -- OpenStack is their control plane and provides the role based controls across the organization. Over 20,000 software releases have been cycled through their new, OpenStack-based platform over the past six months.
Wells Fargo is just one example, but as Foundation COO Mark Collier said in his keynote, OpenStack is being used by the biggest companies in different sectors, companies that have billions of customers, and millions of employees. That's no longer the realm of a mere science project.