Massively Scalable Platform for Data Farming Supporting Heterogeneous Infrastructure

Data farming is a scientific methodology, which heavily depends on technical advances in high throughput computing to generate large amounts of data with computer simulation to investigate studied phenomena. Unfortunately, the availability of versatile data farming systems is very limited and none of existing tool enables integration with novel Cloud solutions. This paper presents a flexible platform for conducting large-scale data farming experiments on heterogenous computational infrastructure including: clusters, Grids and Clouds. Another important feature of the presented platform is the support of interactive data farming experiments, which includes an online analysis of partial experiment results and experiment extending capabilities.