Why CycleCloud?

Why use external Clouds for HTC/HPC?

Before Cloud Computing, when buying a cluster, end-users would size it to be able to complete their largest set of calculations in a desired time. On the storage side, the same sizing would occur to ensure that enough space existed to hold the working data and final results of calculations.

Purchasing a cluster required large up-front capital expenditures for the machines and filers required to do calculations, as well as lengthy procurement and provisioning processes. So when a cluster is operational for the first time, it isn't full to capacity and researchers only have a fixed-sized cluster to do their calculations.

For example, a 4,000 compute-hour calculation takes 100 hours to run on a 40-core cluster. On CycleCloud, the math is different: having 40 processors work for 100 hours costs the same as having 1,000 processors run 4 hours. Results come back the same day rather than four days later.

With CycleCloud™, results come back tens of times quicker at no additional cost.

Features & Benefits

Many Fortune 500s have used CycleCloud to run calculations in EC2, leveraging these key features:

• Automatically start large Condor pools for grid computing, built with Amazon Web Services
• Leverage the set of ready to use applications for Life Sciences, Insurance, and Finance to get started quickly
• Our clusters automatically change size depending upon the workloads, system failures, and cost constraints for calculations
• Execute calculations knowing your software and data are secure with various forms of encryption and randomly generated key management
• Easily manage your cluster with the same CycleServer™ management console that is used by Fortune 500 companies
• Store and use large datasets with ease, remotely mounting your filesystem from your environment
• Use RESTful interfaces to submit/manage data and calculations to a cluster
• Remove the need to purchase a cluster, as well as enabling bursting calculations outside of internal resources
• Get up and running with 100 node clusters without the need for IT staff, datacenter space, or upfront equipment costs