Support for bioinformatics applications through volunteer and scalable computing frameworks

Many e-science applications can benefit from the elasticity of resources provided by volunteer and distributed platforms. While the former is based on resources assigned voluntarily by its owners, the second is based on resources specially configured for this purpose. In this paper, we present the integration of BOINC and Hadoop, two of the most popular middlewares available today, at the infrastructure level of mc², a platform designed to support scientific applications. We discuss some case studies related to DNA sequencing running in Cloud-Burst, model selection running in the MEGA Computational Core and similarity searches running in BLAST. Our results show scientific users can run their experiments in our platform with a high level of abstraction and, in some cases, with very good performance.