OpenStack, an open-source project that early backers once thought could provide an alternative to Amazon Web Services, has fallen decidedly short of that goal. Still, the foundation behind the development of the cloud computing platform thinks the software might find new life as edge computing starts toe emerge.
Mark Collier, an early leader of the project and now chief operating officer of The OpenStack Foundation, acknowledged the reality of cloud infrastructure computing in 2017 in a recent conversation. The public cloud has gone from an experiment to a part of almost everyone’s IT strategy, but Amazon Web Services dominates this market. Microsoft is a strong second, and there is a pretty crowded group of others below them fighting for relevance.
First introduced in 2010 by RackSpace and NASA, OpenStack was intended to be an open-source way for enterprise tech companies to provide something competitive with AWS through a collection of services, including compute, storage, networking and a lot of other cloud essentials. After years of its vendor partners running amok scrambling to catch up to AWS, buyers took their business elsewhere, and the project settled into a role as the go-to product for private clouds — cloud-native philosophies put into practice on self-managed infrastructure — and the basis for several niche public cloud providers around the world.
Half of the Fortune 100 companies are running OpenStack, Collier said. Because anyone can download OpenStack and set it up on their own, “we don’t really have precise numbers, but we certainly believe there are many, many thousands of OpenStack clouds out there,” he said.
The project has caught on inside the media and retail industries, he said, and CERN is also using it to power an internal cloud that manages the Large Hadron Collider. AT&T uses it for its DirecTV streaming service, and Walmart is actually running its online store on OpenStack, Collier said.
While it’s pretty clear OpenStack isn’t going to be an AWS killer any time soon, there are new places emerging where Collier thinks OpenStack might make sense.
Edge computing is one of them. As more large multinational companies start to realize the size and quality of the data they can get from placing sensors around their assets, they’re going to buy more of those sensors. And over time a lot of people — including Microsoft CEO Satya Nadella — think that those sensors will need to more and more computing themselves because of latency concerns in sending that data back to a centralized cloud provider.
“I think that the notion that (workloads are) just going to go to three or four massive companies that’s just going to run all the infrastructure in some centralized location is not going to come true,” Collier said. The big cloud providers are certainly aware of the edge computing trend, and will deploy services around it to try and grab that business, but services built around OpenStack at least have a chance of competing for that business before the market matures, he said.
There’s also still some opportunity among companies that would like to adopt a hybrid cloud strategy but realize their on-premises infrastructure could use a face lift.
“I think we kind of reached this point where people know that in order to really manage infrastructure at scale, you’ve got to do it through software,” Collier said. “I think OpenStack has come a long way, in terms of being reliable and scalable.”
OpenStack is something of a cautionary tale among cloud computing veterans, and it’s illustrative of challenges faced by industry standards groups.
The OpenStack Foundation tried to be Switzerland among its base of tech vendors when it came to how the software would actually run in data centers, and that resulted in vendors like Rackspace, IBM, and Red Hat competing fiercely to offer the “best” OpenStack implementation. Several members of the Cloud Native Computing Foundation, which is driving the conversation around modern cloud computing practices, made it very clear to me earlier this year that they had no intention of repeating the mistakes OpenStack made in trying to manage the demands of tech vendors.
Collier acknowledges that the project might do a few things differently if it could hit refresh, but, as you might expect, isn’t ready to give up on OpenStack.
“We’re basically a collection of people — 85,000 members of our foundation, thousands of developers in 180 countries — that are trying to build open infrastructure so people have an alternative for when it makes sense,” he said. If emerging technologies like containers and serverless really do allow users to move their workloads between clouds, Collier thinks OpenStack could play a role as one of those destinations.