The OpenStack community is infighting. Again. This time, as ever, some of OpenStack's most prominent advocates are arguing over what interoperability means, and just who measures up. Ironically, what OpenStack needs to succeed has less to do with interoperable implementations of the OpenStack reference architecture and more to do with a single vendor imposing standardization on the community.
That's right. OpenStack needs a hegemon.
Every Community Needs A Leader
After all, no one talks about Amazon Web Services interoperability. Or, rather, we don't ask Amazon to sit in committee rooms and conform its API to OpenStack, Cloudstack and other cloud offerings. Amazon dominates public clouds. It sets the standard and others either conform to its API or not. (At present, the OpenStack community continues to vote for "not," which leads industry pundits like Simon Wardley to dismiss it as a "dead duck.")
This is why Cloudscaling CTO and OpenStack luminary Randy Bias is spot on to deride OpenStack interoperability as key to its success:
The very idea that OpenStack requires interoperability to succeed is completely laughable. Interop doesn’t guarantee success.— Randy Bias (@randybias) July 23, 2013
Far too often in the open-source community, we ascribe magical powers to committee-fostered interoperabilty, when the industry actually displays a remarkable tendency toward de facto standardization.
Linux, for example, really took off when the industry was able to rally around a single vendor, Red Hat, as its standard bearer. Perhaps it's not surprising, then, that Red Hat executive vice president Paul Cormier declares that "OpenStack is Linux all over again," with a need for OpenStack's disparate elements to cohere into a distribution that makes it consummable by the masses.
Red Hat At The Helm?
It's not surprising that Cormier sees his company as the unifying center of the OpenStack universe. While there still is a long way to go in establishing OpenStack leadership, Red Hat currently sits in prime position to lead OpenStack just as it led Linux. Why? Because it contributes the most to the project:
In open source, being the source of code matters more than source code. She who contributes most, controls (or influences) most.
Hence, even as the OpenStack community mulls over the introduction of RefStack, a proposed conformance testing system to help evaluate just how "pure" an OpenStack implementation is, OpenStack's members are already jockeying to be RefStack. If Red Hat or another vendor establishes primacy in the community, its implementation will essentially become "RefStack."
Again, Linux illustrates how this will happen. As Red Hat established the lead in the early 2000s, the rest of the Linux vendors, led by Caldera, attempted to band together as United Linux. I was involved in these discussions and can report with some accuracy that they completely, totally failed. It turns out the market didn't want a common Linux distribution created by committee. They wanted the industry standard, which happened to be Red Hat.
Choose Your Hegemon
There's still a long way to go to determine who will lead OpenStack, though Red Hat currently occupies pole position. But whether Red Hat or Rackspace (which used to dominate) or IBM or another vendor, OpenStack's success depends upon a single large vendor assuming leadership. Interoperability will follow, but it won't be designed by committee. Rather, it will be imposed by a dominant OpenStack vendor, with the community necessarily conforming to its distribution.