CloudSpan: Enabling Scientific Computing Across Cloud and Grid Platforms

Within the Nuclear Physics community, grid computing has been established as paradigm for data sharing and computational analysis on a massive scale. In addition, cloud computing has recently emerged as a paradigm in which users lease the resources required to maintain and create virtual storage and computational elements in a shared hosting environment. This project will develop an infrastructure that allows users to execute their scientific applications seamlessly in either environment. The approach is to develop (1) a scheduling system that can reason about the actions required to efficiently place applications on either the grid or cloud platforms, and (2) an analytics service that learns models of how cloud resources behave and predicts how jobs will run on those platforms. Commercial Applications and other Benefits as described by the awardeeCloud computing has emerged as a paradigm used by most online applications including social networking and picture sharing. Beyond the scientific market, the technology should enable the consumer market to run applications requiring intense computing resources, such as video editing and rendering