RightScale Cuts Own Cloud Costs by Switching to Docker

The content below is taken from the original (RightScale Cuts Own Cloud Costs by Switching to Docker), to continue reading please visit the site. Remember to respect the Author & Copyright.

Less than two months ago, the engineering team behind the cloud management platform RackScale kicked off a project to rethink the entire infrastructure its services run on. They decided to package as much of its backend as possible in Docker containers, the method of deploying software whose popularity spiked over the last couple of years, becoming one of the most talked about technology shifts in IT.

It took the team seven weeks to complete most of the project, and Tom Miller, RightScale’s VP of engineering, declared the project a success in a blog post Tuesday, saying they achieved both goals they had set out to achieve: reduced cost and accelerated development.

There are two Dockers. There is the Docker container, which is a standard, open source way to package a piece of software in a filesystem with everything that piece of software needs to run: code, runtime, system tools, system libraries, etc. There is also Docker Inc., the company that created the open source technology and that has built a substantial set of tools for developers and IT teams to build, test, and deploy applications using Docker containers.

In the sense that a container can contain an application that can be moved from one host to another, Docker containers are similar to VMs. Docker argues that they are a more efficient, lighter-weight way to package software than VMs, since each VM has its own OS instance, while Docker runs on top of a single OS, and countless individual containers can be spun up in that single environment.

Another big advantage of containers is portability. Because containers are standardized and contain everything the application needs to run, they can reportedly be easily moved from server to server, VM to VM (they can and do run in VMs), cloud to cloud, server to laptop, etc.

Google uses a technology similar to Docker containers to power its services, and many of the world’s largest enterprises have been evaluating and adopting containers since Docker came on the scene about two years ago.

Read more: Docker CEO: Docker’s Impact on Data Center Industry Will Be Huge

RightScale offers a Software-as-a-Service application that helps users manage their cloud resources. It supports all major cloud providers, including Amazon, Microsoft, Google, Rackspace, and IBM SoftLayer, and key private cloud platforms, such as VMware vSphere, OpenStack, and Apache CloudStack.

Its entire platform consists of 52 services that used to run on 1,028 cloud instances. Over the past seven weeks, the engineering team containerized 48 of those services in an initiative they dubbed “Project Sherpa.”

They only migrated 670 cloud instances to Docker containers. That’s how many instances ran dynamic apps. Static apps – things like SQL databases, Cassandra rings, MogoDB clusters Redis, Memcached, etc. – wouldn’t benefit much from switching to containers, Miller wrote.

The instances running static apps now support containers running dynamic apps in a hybrid environment. “We believe that this will be a common model for many companies that are using Docker because some components (such as storage systems) may not always benefit from containerization and may even incur a performance or maintenance penalty​ if containerized,” he wrote.

As a result the number of cloud instances running dynamic apps was reduced by 55 percent and the cloud infrastructure costs of running those apps came down by 53 percent on average.

RightScale has also already noticed an improvement in development speed. Standardization and portability containers offer help developers with debugging, working on applications they have no experience with, and flexibility in accessing integration systems. Product managers can check out features that are being developed without getting developers involved.

“There are certainly more improvements that we will make in our use of Docker, but we would definitely consider Project Sherpa a success based on the early results, Miller wrote.