TLA 2013: IT Management
CERN, the European organization for nuclear research, operates the world's largest particle physics laboratory, hosting 10,000 visiting scientists and engineers representing 608 universities and research facilities.
You've likely read about its latest discovery, a particle believed to be the Higgs boson -- a discovery in the science community equivalent to landing a man on the moon. The Higgs boson search was conducted through CERN's Large Hadron Collider, which is the world's most powerful particle-smasher. This collider produces 1 petabyte of data per second and demands an astronomical rate of computing power. But it's not the only collider at CERN producing data.
The growth in CERN's computing capacity forced the team to create a new data center similar in size to its existing facility, which held more than 10,300 servers, 83,000 disks, and 90,000 cores. However, with a limited budget, the IT staff at CERN was not allotted new staff members to help manage this new data center, which by 2015 will amount to more than 15,000 hypervisors running 150,000 virtual machines.
Bell and his team needed to come up with a cost-effective way to manage both of these data centers with innovative technologies and methodologies. For years, the team had been building its own software because it was on the bleeding edge of technology (for example, the World Wide Web was invented at CERN in the 1990s to address the challenges of sharing information). But continuing to create hardware and software from scratch with the amount of staff available was not feasible.
The big breakthrough came when the team noticed that industries in Web and cloud were facing a similar problem to CERN. Although CERN's data center was specialized, it could still use a similar blueprint. The team used the new OpenStack cloud-based data center standard and an automation tool called Puppet.
Through these technologies, CERN was able to create its infrastructure to scale in a matter of months -- not years -- without having to customize every piece of hardware and software. As a result, Bell's team was able to stay on budget, scale CERN's infrastructure more quickly, and create a system to address CERN's future infrastructure.