Investigating cloud computing as a solution for analyzing particle physics data
Abstract
The issue of research data preservation is under discussion around the world. In High Energy Physics (HEP) this has recently become an important topic as a number of experimental projects have stopped recording collision data and there is a desire to ensure that the data is useable for many years. Although the HEP community are pioneers in scholarly communication, the community has taken yet few steps to enable re-use of its data. A solution will likely involve a variety of technological tools such as virtualization, grid and cloud computing. The goal of this project is to build an environment for the analysis of legacy HEP data with particular emphasis on the BaBar experiment.
Intellectual Merit
Preserving the ability to analyze BaBar data requires that we also maintain the software and the calibration databases. One potential solution to this problem is through the use of computer virtualization technologies. Virtualization will allow one to encapsulate the BaBar software in a supported operating system making it relatively independent of the local machine configuration. Initial studies using Xen virtualization software andthe BaBar software are underway.
Broader Impact
The outcome of the project will be a facility that will enable the BaBar collaboration to analyze its data for many years. We see it being a model for the long-term persistence of data from other HEP projects and possibly other fields of science. We believe that future HEP projects need to consider their long-term data strategy in the initial design of the experiment. There is a proposal for a next generation BaBar experiment called SuperB to besituated in Italy.
Use of FutureGrid
We wish to investigate the use of clouds for generation of simulated and the analysis of real data recorded by the BaBar Experiment. The group at the University of Victoria are members of the BaBar Collaboration based at the SLAC National Laboratory on the Stanford University campus.
Scale Of Use
We wish to use 10s of VMs on a periodic basis. We occasionally may wish to run around 100 VMs.
Futuregrid is a resource provider for