GE

How Agentless Architecture Saved GE 25 Hours per Deployment

Download the Case Study


The Environment

  • XL Deploy
  • XL Release
  • Jenkins
  • CA Agile (Rally)
  • Flowdock
  • Tableau

The Challenge

GE Power Fleet Services’ development and production teams had hit a wall in their strive towards continuous improvement. Unplanned work interrupted planned work, software release processes moved too slowly to meet business demands, and painful, disjointed handoffs between teams disrupted the release process. They wanted to develop higher quality software faster and improve collaboration between development and production. To fully realize the benefits of their move from waterfall to Agile methods, GE Power knew they needed release automation.

The Solution

GE Power chose the XebiaLabs DevOps Platform to automate their deployments and to orchestrate and control their release pipelines. A major factor in their decision was XebiaLabs’ agentless architecture.

Results

  • Releases that took months now take days and only 1/3 of the resources
  • 25 hours per deployment saved, multiplied by hundreds of deployments per year
  • Higher quality software – deployments go into production “first time right”
  • Elimination of rework increased capacity to innovate, which boosted revenue growth
  • Almost 2X return on their XebiaLabs investment this year and expect 5X return next year
  • Critical release data helps team make quick, data-driven decisions and measure success
  • Migration from legacy process to release automation was accelerated due to XebiaLabs’ ease of use

Background

The GE Power Fleet Services team develops software to remotely monitor, run diagnostics on, and capture analytics for the industrial Internet solutions they deliver. They wanted to move from waterfall to an Agile approach. After improving the front end of their software development process with Continuous Integration using Jenkins, they realized they had a lot of software sitting on the shelf waiting to be released.

They needed to automate their deployments and orchestrate their entire release pipeline in order to reach their goals of developing software faster, better, and less expensively. Other departments throughout GE Power had the same problem, so the team wanted to create a repeatable approach throughout the organization. To streamline and standardize their DevOps processes, GE Power evaluated IBM, CA, and XebiaLabs for release pipeline orchestration and deployment automation.

Requirements

  • Standardize, accelerate, and automate deployments to Dev, Test, QA, Staging, and Production environments  
  • Orchestrate, manage, and standardize release pipelines across all tools in the environment
  • Build once, deploy many: push identical packages through multi-stage automated pipeline
  • Collect release details 
  • Visualize the release data for stakeholders 
  • Build on their Continuous Integration foundation
  • Integrate with tools in the environment, such as CA Agile Central (Rally) and Flowdock chat

Developers Deploy Twice as Fast  

After implementing Continuous Integration, GE Power chose XL Deploy to automate their deployments as their next step. XL Deploy was a clear choice over its competitors due to its agentless architecture. Introducing a new technology into the stack with the infrastructure team was typically a lengthy process. But with agentless technology, they could easily deploy their applications into the stack because they were not required to install agents. As a result, there was no interruption to the infrastructure team.

Automating their deployments solved GE Power’s initial bottleneck. Defining and improving their pipeline between development, QA, and production was their next DevOps challenge. Early phases of the pipeline were not representative of the production environment: for example, the configurations were often different, which led to unpleasant surprises in production. As they found defects or failures in the pipeline, they often had to abort entire releases, which meant losing critical release metrics.

To meet this challenge, GE Power created a controlled pipeline with XL Release. The tool provides the orchestration they need in order to define their processes and understand where to automate and how to standardize each stage of the pipeline.

GE Power is also able to drive continuous improvement with XebiaLabs using XL Release’s metrics and dashboards. XL Release provides metrics that show release failures, identify problems, and provide information to help them solve problems faster.

"We automated over 20 packages and found we saved over 25 hours per deployment. One analytic application that took months to get through the pipeline now takes us only days with XebiaLabs. The team is finally deploying into production ‘first time right,’ with no failures."

- Eric Haynes, Director of Architecture and Engineering, GE Power Fleet Services

Now that GE Power has XebiaLabs in place, they can roll back a specific phase of the release if they find an issue, fix the problem, and maintain the overall release metrics so that no critical data is lost. XL Release provides auditability and traceability so that continuity of release metrics is maintained.

They further benefit from XebiaLabs’ detailed reporting and metrics by plugging their release data into Tableau to customize their reports. These reports enable their development and production teams to collaborate in ways they never have.

Reaching Returns in Year One

GE Power realized a 2X return on their release automation investment in the first year. Releases that took months to get through the pipeline now take days. “We automated over 20 packages and found we saved over 25 hours per deployment. One analytic application that took months to get through the pipeline now takes us only days with XebiaLabs. The team is finally deploying into production ‘first time right,’ with no failures,” noted Eric Haynes, Director of Architecture and Engineering, GE Power Fleet Services.

Now that GE Power has a standard approach, streamlined release pipelines, and automated and predictable outcomes, their DevOps and scrum teams can collaborate easily – one of the main goals they set out to accomplish. With XebiaLabs, they eliminated release and deployment bottlenecks and minimized errors, so they now produce higher quality software releases.

GE Power realized a 2X return on their release automation investment in the first year.