SHARE

Ubuntu’s Shuttleworth Explains Why Not All Containers are the Same

Mark Shuttleworth, the founder of Ubuntu Linux, was an early backer of OpenStack as well as containers. This week, Shuttleworth’s company Canonical announced new commercial support for Kubernetes, which is a widely deployed container orchestration and management engine. In an interview with Datamation Shuttleworth emphasized that it’s important to understand the different use cases for […]

Written By

SMK

Sean Michael Kerner

Sep 29, 2016

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More

Mark Shuttleworth, the founder of Ubuntu Linux, was an early backer of OpenStack as well as containers. This week, Shuttleworth’s company Canonical announced new commercial support for Kubernetes, which is a widely deployed container orchestration and management engine.

In an interview with Datamation Shuttleworth emphasized that it’s important to understand the different use cases for containers and what the different types of container systems are all about.

“There are going to be different types of container co-ordination systems,” Shuttleworth said. “There will trucks, tractors and cars.”

Shuttleworth emphasized that in his view, he doesn’t expect there to only be one way to manage container processes. For Shuttleworth, the Docker world is all about co-ordinating processes, while the OpenStack world is about co-ordinating machines.

“Kubernetes is really stepping up to lead the hyper-elastic, on-demand, operational story,” Shuttleworth said.

With Kubernetes the story is also about stateless applications, essentially things that are dealing with single process requests, or requests related to a specific customer. Shuttleworth expect that databases will run outside of Kubernetes control and that Kubernetes will focus on the hyper-elastic application code.

“Kubernetes will probably become the leading way to handle Docker-style process container, based on the level of excitement we’ve already seen,” Shuttleworth said. “But there is still room for Mesos, Docker Datacenter and the right strategy for us (Canonical) is to offer our customers a fully supported stack for each of those things.”

Shuttleworth added that a lot of people that are running Ubuntu in the enterprise are using Kubernetes already. As such, he commented that what Canonical is now doing is connecting the dots between what people are trying out, to mechanisms for enterprise support.

In 2015, the OpenStack Magnum effort got underway as a way to manage containers inside of OpenStack. Rackspace uses Magnum today as the means to enable its Carina Containers as a Service platform. Shuttleworth isn’t particularly a fan of OpenStack Magnum.

“People use the word containers today to describe very different things,” Shuttleworth said.

He explained then when a user uses Docker to run a MySQL database for example, Docker is being used to launch a binary. That binary is confined and isolated from the rest of the system, bringing its own filesystem. In Shuttleworth’s view Docker containers are largely about running processes.

“At the end of the day, no one wants a process co-ordination system that works on one substrate,” Shuttleworth said. “Why would you only invest in a process co-ordination system that works on OpenStack when you know you need to operate across multiple public clouds.”

Understanding that context is why Canonical has backed Kubernetes and is now supporting it .

“Our Kubernetes offering is pure Kubernetes and we we’re enabling people to get it in the same form on Google Cloud, Azure, OpenStack, VMware and bare metal,” Shuttleworth said. “That resonates more directly with what people want.”

Sean Michael Kerner is a senior editor at Datamation and InternetNews.com. Follow him on Twitter @TechJournalist