Cloud Computing as a Source of Independent, Large Scale Computing Capabilities for Small Science Teams, Based on Pierre Auger Collaboration Example

Small science teams often have to conduct researches that needs heavy calculations without access to adequate processing powers of computation clusters. However, cloud computing can be used as a reliable source of calculation potential. Development of software for Pierre Auger Surface Detectors needs large amounts of data, that can be obtained only by simulations conducted by CORSIKA and OffLine packages. With help of scripts, it is easy to create a nearly fully automated, cost and time efficient system, that can be used to many different simulation cases. Detailed instruction is presented, along with solutions of problems encountered during the simulations. Architecture of the system was designed in such way, that minimal configuration changes are needed to develop new simulation cases and they are provided as outside module. As an example, comparison of calculations from Lodz team will be presented as a proof of efficiency of the system, that will be later presented for the whole Pierre Auger Collaboration, to help other research teams.

I. Problem of Access to Huge Calcualtion Power in Research

Modern research is based on large pool of data, both from experiments and simulations. Some fields of knowledge are even more dependent on it than others. However not all science teams have enough computing power to perform large scale calculations. University HPC clusters needs to be shared with other users and the tasks are strictly scheduled. Also, not every form of simulations are suited to be used in that kind of machines. Some educational facilities have no such hardware, or have need to perform large scale calculation too seldom, to justify the costs of building and maintaining such equipment.