During the last 20 years, computing saw the addition of more and more abstraction layers (virtualization, cloud APIs, application orchestration…). With that evolution we saw the emergence of a new role: the infrastructure provider. Infrastructure providers are tasked with providing, maintaining and evolving the computing, networking and storage infrastructure that others use to deploy their applications and store their data on.
Before 2010, providing infrastructure had traditionally been the realm of proprietary hardware and software. That’s when OpenStack was started, and it changed the game. By proposing an open source solution to build a pure infrastructure as a service cloud, OpenStack bootstrapped the concept of open infrastructure: pure open source solutions for infrastructure providers.
The huge success of OpenStack was soon emulated by other open infrastructure solutions, like Kubernetes. Following a similar open development model, Kubernetes standardized the application deployment layer and gave birth to the cloud-native landscape. But cloud native assumes that there is a cloud underneath – and the way to do it using open source solutions is to combine Kubernetes with OpenStack, to provide the lower level infrastructure it needs. Like Jonathan mentioned, that’s the LOKI stack: Linux, OpenStack, Kubernetes Infrastructure.
There are of course numerous other software components in the OpenInfra landscape. Some of them were created after OpenStack, like Kata Containers, allowing infrastructure providers to offer secure container runtimes to their users. Some others, like Ceph, were created before, but OpenStack gave them the exposure they deserved within an OpenInfra stack.
That is what the past decade looked like. A time for pioneers to take hold in a space where only proprietary solutions existed before. What will the next decade look like? All signs point to more usage of open source in infrastructure: more software components, deployed everywhere and at all scales.
If 2021 is any guide, we observed a dramatic increase in the number of deployments reported in our OpenStack user survey: +66% growth year-over-year. We now count 25 million CPU cores managed by OpenStack. That increase comes from existing users dramatically increasing their footprint, with users like Walmart or LINE running over 1 million cores each and China Mobile running over 6 million. But it is also coming from lots of new users.
What are the trends driving this increase in OpenInfra adoption? The first one is the rise of digital sovereignty concerns, as lots of countries (and companies) are waking up to the vulnerability of handing over all their infrastructure needs to potential adversaries. This explains the recent increase in the number of public clouds powered by OpenStack, but it also explains the appetite for more open source solutions for providing infrastructure, beyond OpenStack.
The second trend is niche players pursuing innovating business models. Open infrastructure makes that permissionless innovation possible. Startups like OneQode or Exaion provide infrastructure for very specific use cases, leveraging open infrastructure solutions. That is my favorite aspect of open infrastructure: enabling everyone to play with this technology and innovate.
Ten years in, open infrastructure has grown beyond the pioneers, and it’s here to stay. At the OpenInfra Foundation, we want to nurture this movement and help shape what the future of open infrastructure looks like. We are just getting started. Join us!
Read the full 2021 OpenInfra Annual Report here!