CERN didn’t have a single solution that could handle several different projects and use a single database, ensuring the deployment could expand easily. The various options they explored exhibited varying limitations including: an inclination for only one programming language, the lack of an API and insufficient scalability or speed for the job. Some were limited in terms of platforms they would work with, or proxy features. They needed scalability and flexibility all in one.
How did CloudBees Help?
CERN’s ROOT team with ROOT being their open source foundation product on a global scale developed for high volume data analysis, implemented CloudBees to automate its build and test processes. The builds and tests are now running fast and efficiently. Prior to implementing CloudBees, frequent builds and tests were not regularly performed because of slow processes and inefficiencies.
- Complete visibility. Nightly binary snapshots are created so the entire team receives user confirmation the next day. This allows for a “blessed,” tested binary snapshot to be released daily.
- Process automation. CloudBees Flow™ automated CERN’s build and test processes. Code is developed with Subversion, which ports into CloudBees Flow. The builds are run using a continuous integration process, and the test suite is then invoked every 2 to 3 hours with CloudBees Flow.
- Streamlined workflow efficiency. CloudBees Flow has dropped directly into CERN’s existing development workflow, allowing them to be more efficient and reliable with a single stroke. The success the ROOT team has seen has encouraged other teams at CERN to use CloudBees Flow. There are now four distinct projects at CERN using CloudBees Flow every night, with each of these projects building on multiple platforms, and this list of projects and teams adopting CloudBees continues to grow.
- Single scalable solution. CloudBees Flow was able to accommodate for production-grade, multi-platform, multi-version systems designed to scale as needed.