How GE Healthcare migrated 500 applications to the cloud

GE Healthcare recently migrated its infrastructure to cloud — and ‘remotely managed private cloud’ was a big part of that migration.

Patrick Weeks
Swapnil Bhartiya

At the OpenStack Summit, Jonathan Bryce, the executive director of the OpenStack Foundation talked about a new consumption model ‘remotely managed private cloud’ that’s gaining traction. GE Healthcare is already using that model through Rackspace.

During his keynote speech, Patrick Weeks, senior director of cloud, automation and operations at GE Healthcare (pictured above), detailed the work that the company has done over the last couple of years to move its healthcare enterprise applications to public cloud and remotely managed private cloud.

GE Healthcare’s environment was a good example of the ‘3Cs’ that Bryce mentioned in his keynote speech.

GE Healthcare is a highly regulated environment, which also tends to be very complex as it has grown through multiple acquisitions. They had to deal with server sprawl, redundant applications, and legacy applications that may be on the edge of their support cycle.

While GE Healthcare has a massive infrastructure, it is not an infrastructure company. It had to approach the problem in a more strategic manner. The company came up with a three-pronged strategy:

Commitment: Success of the projects required commitment throughout the organization — from the IT leadership, business leadership, the engineers who were going to migrate the applications and from the end users who were going to be testing it.

Involvement: Everyone who had a role to play in this transformation was brought to the table, even before they got started. Weeks stressed getting the security and compliance people onboard from very early on as these are the people whose permission you would need in the end, due to the regulatory requirements. They got a set of problem solvers at the table who would help them find solutions as they came across problems.

Impossible Goals: The third and the last strategy was to set really impossible goals. “Without an impossible goal, you will not drive the creativity necessary. You will not build the capabilities necessary to go to the next level,” said Weeks. The impossible that they set was moving 126 applications from their infrastructure to the cloud in the first year.

With this strategic planning, GE Healthcare ended up migrating some 500 applications to the cloud that represented over 42 percent of its business. They purged, or retired, some 600 applications as they moved to the cloud, which translated into roughly $30 million in annualized savings.

While public cloud was at the center of this migration, private cloud was a critical piece of the equation for various reasons.

GE Healthcare needed to be able to access its internal applications. The company could have done that with public cloud, but that makes things extremely complicated, according to Weeks. On top of adding complexity, Weeks argued, what’s the point of spending time, money and effort trying to solve some of those challenges when you already have a better solution? “Put it in private hosted cloud and simplify your journey,” he said.

And when you consider security and compliance, nothing beats a private hosted single tenant environment to host your applications and offer the most extreme level of security for your most sensitive data.

The company had already done a lot of work on public cloud with their migration and  didn’t want to reinvent the wheel with private cloud. It wanted to leverage platforms across the clouds it was using. It created GEIX (GE Infrastructure Xchange), a remotely managed private cloud-as-a-service that was powered by OpenStack and offered through Rackspace.

Rackspace helped get GEIX ready in November last year and GE Healthcare started to migrate some of its development environments to it. Once the engineers successfully performed all the testing and were satisfied with the results, they moved into production just eight days later.

What usually takes weeks and months for a new platform was accomplished in a matter of hours and days. They managed to move 70 applications within the next six months.

Weeks attributed the success of this migration to creativity and flexibility of OpenStack. “OpenStack allowed us to tailor our solution,” said Weeks, “Cookie cutter ways of being able to deliver our infrastructure.”

Last but not least, Weeks highlighted the most critical aspect of using OpenStack: open source and no vendor lock-in.

“Haven't we paid enough to our primary partners over the years?” said Weeks. “If we're going to have a competitive IT solution for the future, we need to embrace platforms such as this.”

This article is published as part of the IDG Contributor Network. Want to Join?

NEW! Download the Spring 2018 digital edition of CIO magazine