What's new in OpenStack?
What's new in OpenStack?
The world of open infrastructure is constantly changing. Learn how OpenStack is pivoting to meet the needs of its user community.
The OpenStack global community is gathering together in Vancouver, British Columbia this week to collaborate, learn, and build the future of open source cloud computing.
As OpenStack Foundation Chief Operating Officer Mark Collier referenced in his opening keynote, the uses which OpenStack is seeing today expand far beyond what most who were involved in the early days of the project could have ever imagined. While OpenStack started out primarily in the traditional data center and found many large-scale users, particularly in the telecommunications industry, who were using it to manage huge installations of traditional x86 server hardware, the flexibility of OpenStack has today allowed it to thrive in many other environments and use cases.
Today, we see OpenStack powering everything from academic and research projects to media and gaming services, from online retail and e-commerce to manufacturing and industrial applications, and from finance to healthcare. OpenStack is found in all of these different places not just because it is cheaper than using the public cloud, not just because it makes compliance with various regulations easier, but because its open source code makes it flexible to all sort of different situations.
For example, it can run applications like machine learning and artificial intelligence that make use of graphics processing units (GPUs) instead of traditional CPUs to power their performance, and divide those GPUs across multiple applications running inside of virtual machines. It can allow for deployment to the diverse locations around the globe where scientific research takes place, which generates orders of magnitude more data than could ever be captured and uploaded to the public cloud. But perhaps most importantly, it allows organizations to mix and match the software they use to meet their actual needs, rather than relying on a single vendor’s proprietary solutions.
One change we’ve seen in recent years within the OpenStack community is a shift in focus away from the infrastructure itself to empowering the applications which it supports. To that end, two topics getting much attention at this summit are containers, which enable faster and denser deployments of applications, and continuous integration and development, which allow faster development cycles between coding and production.
While containers aren’t new to the OpenStack ecosystem, it seems as if every summit introduces new ways in which container technologies are interacting with OpenStack; either on top, beneath, or alongside OpenStack’s services.
The big news from yesterday was the announcement of Kata Containers hitting its 1.0 release. As a refresher, Kata Containers is a new project within the OpenStack Foundation aiming to build lightweight virtual machines around containers to allow applications to function with the speed and composability of containers while maintaining the security and isolation of virtual machines.
While still a relatively new player, they are already designed to plug in seamlessly with existing projects, including Kubernetes, by adhering to the Open Container Initiative specification.
A relatively new feature of the OpenStack community is the addition of the developer-focused OpenDev event, which this year is focused on continuous integration and development within the context of OpenStack.
While there are a number of CI/CD systems which might fit into the workflow of an OpenStack user, the big focus of this event was Zuul, whose tagline “stop merging broken code” is probably something many organizations can identify with.
Zuul isn’t new, but its place in the OpenStack umbrella is new. Originally developed as the CI/CD system for OpenStack, it now sits as a stand-alone project which can be used as a CI/CD system for any large project which spans multiple repositories or has many external dependencies which may be seeing development concurrently.
In OpenStack, projects depend on one another to function properly. In order to properly test code, you need to know not just that it doesn’t break the project you’re committing code to, but also that it won’t break any of the other projects that depend on the project you’re working on. By better coordinating across projects, Zuul’s goal is to go beyond simple code testing and instead focus on enabling cross-project collaboration and coordination.
You may have noticed that several of the topics coming out of this OpenStack Summit don't fit neatly under the umbrella of what had been previously called OpenStack. They go far beyond just provisioning compute, network, and storage resources.
That's by design.
OpenStack Summit, and the OpenStack Foundation supporting it, have broadened their focus to include sessions on many of the other projects, technologies, and other things that modern IT operations professionals need to be familiar with. In many ways, this represents a shift from a focus on software to a focus on people and what those people need to be able to do their jobs.
To this end, in addition to 250+ talks on OpenStack, the Vancouver agenda included 70+ Kubernetes sessions, as well as talks on other technologies like Docker, OPNFV, ONAP, Ceph, Ansible, Tensorflow, and many others.
It’s an exciting time to be working in open infrastructure!