A Cost Benefit Study of Doing Astrophysics On The Cloud: Production of Image Mosaics
Bruce Berriman (IPAC, Caltech)
Bruce Berriman (IPAC, Caltech), Ewa Deelman (ISI, USC), Gurmeet Singh (ISI, USC), Miron Livny (University of Wisconsin, Madison), John Good (IPAC, Caltech)
Utility grids such as the Amazon EC2 and Amazon S3 clouds offer computational and storage resources that can be used on-demand for a fee by compute and data intensive applications. The cost of running an application on such a cloud depends on the compute, storage and communication resources it will provision and consume. Different execution plans of the same application may result in significantly different costs. We studied via simulation the cost performance trade-offs of different execution and resource provisioning plan by creating, under the Amazon cloud fee structure, mosaics with the Montage image mosaic engine, a widely used data and compute intensive application. Specifically, we studied the cost of building mosaics of 2MASS data that have sizes of 1, 2 and 4 square degrees, and a 2MASS all-sky mosaic. These are examples of mosaics commonly generated by astronomers. We also study these trade-offs in the context of the storage and communication fees of Amazon S3 when used for long-term application data archival. Our results show that by provisioning the right a mount of storage and compute resources cost can be significantly reduced with no significant impacton application performance.
Mode of presentation: oral