If you follow the evolution of OpenStack, you know how it’s finding its way into all sorts of workloads, from high-level research to car manufacturing to all-new 5G networks. Organizations are using it for everything from the mundane to the sublime and sharing what they’re learning with the OpenStack community.
Some of the examples offered up at the recent OpenStack Summit Berlin showed that OpenStack is a full-fledged part of the IT mainstream, which means there are a wealth of ideas out there for your own implementation.
In many cases, the advances of others – including Adobe, AT&T, NASA, Oerlikon, SBAB Bank, Volkswagen, Workday and many other companies and organizations, big and small – are being contributed back to the community for you and others to use. This is a critical part of OpenStack and SUSE OpenStack Cloud, which take the best the community has to offer to improve the platform and how organizations solve problems.
Take Workday, the human resources software-as-a-service vendor, which in 2019 expects to have half of all its production workloads living on the 45 OpenStack private-cloud clusters it’s running in its global data centers. That represents about 4,600 servers, up from just 600 in 2016.
To manage the growing demand for its products, Workday created and now manages about 4,000 immutable VM images that are updated on their own cycles, with new versions of Workday deployed every weekend. That means the company needs to regularly tear down and replace thousands of VMs in a very short time and do it without any downtime.
That scale required automation, and the growing complexity required a new effort to gather data about their clusters and OpenStack controllers. They used Big Panda for incident management and Wavefront for monitoring and analytics, looking for anomalies and problems.
As it turns out, they uncovered some real issues with how they deployed images, and solved those problems by extending the OpenStack Nova API to leverage its caching capability to pre-load big images – what they call image pre-fetching. This enabled them to speed up the image deployments so instead of big images slowing down the restart of thousands of VMs, they could pre-load them and relaunch new VM instances quickly.
They did some ingenious stuff, like enabling Glance to serve up images directly to remote OpenStack controllers, and got help from the community for figuring it out. With OpenStack’s complexity, that openness made their work doable, and in the end, they offered their Nova API work back to the community.
Workday is just one example of the companies taking advantage of the power of OpenStack and the open source community to solve real problems. Check out these and other OpenStack successes – including these 51 things you need to know – from the OpenStack Summit Berlin.