Parallelization and distribution strategies of large bioinformatics requests over the grid

Abstract

This paper focuses on simultaneous scheduling of computation and data replication for life science applications on the grid. We present an adaptive algorithm based on the SRA algorithm (Static Joint Replication and Scheduling) [4] with more dynamicity for the jobs frequencies. The use of a linear program giving a databases mapping on the nodes and a jobs distribution schema, ensures us that our data placement and jobs distribution will be near the optimal solution, as long as the informations about the jobs frequencies are right. We validate our results with large jobs submissions simulations on a realistic platform. © 2008 Springer-Verlag Berlin Heidelberg.