Case Study: Cycle Computing
|
Download The Case Study. |
"We amplified our Chef infrastructure and streamlined configuration scripts to help scalability. We developed an approximate timing architecture which benchmarked and facilitated the increased peak number of nodes and the nodes per unit time. Opscode's Chef enabled us to easily and reliably configure our systems to meet our demanding scale requirements."
Jason Stowe, CEO
The Challenge
Cycle Computing helps customers of all sizes gain fast and efficient access to the kind of supercomputing muscle usually reserved for universities and large research organizations. Cycle's founder, CEO and HPC expert Jason Stowe wanted to test the envelope of cloud computing and take its technology to the next level - spinning up a 10,000- core cluster in the cloud. Stowe teamed up with a prominent life science company who needed supercomputing power to research how proteins bind to each other, and utilized Amazon Web Services to begin provisioning 10,000 cores in an HPC cluster on Amazon's EC2 compute cloud.
The Solution
Admeld's Technical Operations department deployed using a CentOS Linux cluster; Cycle built its 10,000 cores using 1,250 instances with eight cores each. Scaling up roughly 2,000 cores at a time, Cycle enlisted its own CycleCloud software, the Condor scheduling system, and Opscode's Chef cloud infrastructure automation engine. Using Chef, the total time to provision the entire cluster was 45 minutes, and cost approximately $1,000 per hour for a 10,000-core cluster.
The Results
- Set up time significantly reduces from weeks or months to minutes.
- Increased agility and reliability to more efficiently manage platform configuration.
- Improved configuration to increase the scale, to enable creation hundreds of execute nodes at once, without any errors.
- Increased cost savings by eliminating cluster ownership/maintenance costs of building out a 10,000 core cluster infrastructure.
About Cycle Computing
Since 2005, Cycle Computing has delivered proven, secure and flexible high performance computing (HPC) and data solutions, helping clients maximize existing infrastructure and speed computations on desktops, servers, and on-demand in the cloud. Cycle helps customers of all sizes gain fast and efficient access to the kind of supercomputing muscle usually reserved for universities and large research organizations.
www.cyclecomputing.com
About Opscode
Opscode is the leader in cloud infrastructure automation. We help companies of all sizes develop fully automated server infrastructures that scale easily and predictably; can be quickly rebuilt in any environment, and save developers and systems engineers time and money.
Our team is comprised of web infrastructure experts responsible for building and operating some of the world's largest websites and cloud computing platforms. Opscode is headquartered in Seattle, Washington, USA.
