The upcoming OpenShift 3 release will integrate Docker and Google Kubernetes as the basis for Red Hat's cloud platform as a service.
Red Hat is developing a new milestone release of its OpenShift platform-as-a-service (PaaS) technology that will shift the platform to Docker containers and Kubernetes orchestration.
The upcoming OpenShift 3 platform will mark yet another page in the continued evolution for Red Hat's PaaS effort that began with the acquisition of Makara in 2010.
The original Makara code was already turned over when OpenShift 2.x debuted,
and with the upcoming OpenShift 3 release, there is nothing of the original Makara code left, said Joe Fernandes, Red Hat director of product management for OpenShift.
The big shift in the upcoming OpenShift 3 milestone is all about refocusing the PaaS around Docker.
"What OpenShift 3 really represents is what has been happening in the open-source space and Linux containers, driven by Docker and other projects, including Kubernetes and Atomic, that have sprung up around it," Fernandes explained to eWEEK
Docker is an open-source technology for application virtualization using containers that has gained in popularity in the last year. Red Hat's Atomic project is an effort to create an optimized Linux platform for Docker, while Kubernetes is an open-source effort that Google started for container orchestration.
OpenShift is already built around Linux containers, though it had not been leveraging Docker. With Docker, a standard has emerged on how containers can be managed and how applications can be managed to run inside a container, Fernandes said.
Docker alone isn't enough for a full PaaS, and there is a need for an orchestration system to help manage and deploy container applications, Fernandes said. That's where the Google Kubernetes open-source project comes into play.
"From a large-scale orchestration management perspective, Google has more experience than anybody, given the fact that they have run containers at large scale for a long time," Fernandes said.
In the most recent OpenShift 2.2 release, Red Hat is using its own container implementation, called Gears. What runs inside of the Gear is an OpenShift cartridge. While there have been many applications available as OpenShift cartridges, a lot more applications are packaged as Docker images.
"With OpenShift 3, we have essentially eliminated the difference between a Gear and a Cartridge," Ashesh Badani, vice president and general manager of OpenShift at Red Hat, told eWEEK
. "Now it's all just various containers."
A key focus for Red Hat is to enable its users for the hybrid cloud world, where workloads can be deployed on physical or virtual infrastructure as well as on private or public clouds. By using standards like Docker, Badani said that it's easier to deliver on the promise of interoperability across different deployment environments.
One of the biggest benefits of Docker is portability and the fact that the same Docker image can run on Amazon, on OpenStack or bare-metal, Fernandes said.
"Our job in OpenShift is to provide the best environment to run containers," Fernandes said. "That's our challenge, we want to give people the best experience for building out their applications and moving them through the lifecycle."
With the shift away from Gears and Cartridges, Red Hat is building out versions of all applications it had been supporting as Cartridges, in the new Docker image format for OpenShift 3. By embracing Docker, OpenShift users will also be able to pull from other Docker app repositories including the public Docker Hub
. Looking beyond that, Red Hat will also have migration services for customers that need to migrate custom apps that aren't yet available in Docker.
Enterprise application deployments generally require storage and networking capabilities, which is also part of Red Hat's OpenShift 3. Fernandes explained that within Kubernetes, every container gets its own IP address. In terms of how to network multiple containers across different host servers, Kubernetes already has capabilities when running on the Google cloud.
"What we're doing, working with Google and others, is to make sure that the network abstraction works regardless of where you are running Kubernetes or OpenShift," Fernandes said.
Kubernetes also has a model for attaching storage volume to containers. With OpenShift 3, the goal is to be able to enable to map the container storage to enterprise storage technologies.
From a product availability perspective, Red Hat has three streams for OpenShift, including the open-source OpenShift Origin community, OpenShift Online and OpenShift Enterprise products. OpenShift 3 feature development is being done in Origin, and Red Hat already has some of its large customer testing out the technology. Fernandes said he expects there to be a formal beta next month for OpenShift Enterprise.
Sean Michael Kerner is a senior editor at
InternetNews.com. Follow him on Twitter @TechJournalist