OpenStack – four years on. What the experts think went wrong and what went right.
Talking at the OpenStack Summit in Paris today, a roundtable of OpenStack experts illustrated how, after four years of OpenStack, the open source cloud platform is taking the world by storm but is still at the stage of experimentation and innovation.
The speakers also highlighted how the platform is still susceptible to market forces that will change the face of OpenStack over the coming years, and that there is constant pressure to deliver features to the enterprise that is hard to keep contained.
Thierry Carrez, director of engineering at the OpenStack Foundation, said that OpenStack has failed in his initial goal of replacing Amazon public cloud.
Carrez said: "My initial goal was to not have a future where Amazon would provide public cloud. The only thing we could do was to build OpenStack and beat Amazon at its own game. We did accomplish more than what we set out to do but Amazon is still there.
"We don’t yet have this interconnected network of public cloud, we are just not there yet. We took a long detour."
Jay Pipes, principal technical architect at Mirantis, said that his personal goal was to improve the API experience found in the cloud platform.
He said: "I think we have honestly failed to focus on the API side of OpenStack. To this day it’s a mess. Three, four years ago, I was really pushing for the standardisation of APIs, but that hasn’t been seen through."
When discussing what has been the unexpected successes for OpenStack since 2010, Carrez commented that the growth of the platform has been a surprise. Pipes agreed, saying that the growth has been pleasantly surprising.
He said: "We’ve had challenges keeping up with the growth curve. An early criticism was that it’s all fluff and marketing, but I think we’re actually starting to catch up to the marketing hype a bit, which I think is a success. We’re matching what we say OpenStack is."
Monty Taylor, who was working at Rackspace when the OpenStack initiative was first launched, said that no one has really had the chance to create a second iteration of OpenStack, as the first project was picked up so fast that there has been no time to amend the way it is used.
Taylor said: "We had so much marketing hype, then an explosive growth in the code and community, so we never really got the chance to have that second version iteration. We have to keep making this thing work and move forward. Everyone wants it to be so many things and we can’t just take a chunk and throw it away because there are loads of people using it already. How to move that forward in the same way is the ongoing challenge."
All speakers at the Paris roundtable expressed concerns for the challenges OpenStack faces in the immediate future. Bringing up a people issue rather than a technical one, Anne Gentle, who works on OpenStack for Rackspace, said that there are simply not enough females in the project. Gentle said: "I still can’t believe how little women there are."
Pipes put forward the argument that the OpenStack community’s main challenge is learning to say no. He said: "I think that especially with the growth and number of vendors, developers, contributors, operators, and deployers relying on OpenStack in IT, it’s just a never ending stream of feature requests and constant pressure. Whether it’s NFV or IT enterprise architecture, it’s a constant steam, and even if you had a million developers you still couldn’t keep up with it."