SCAN: A Smart Application Platform for Empowering Parallelizations of Big Genomic Data Analysis in Clouds

Abstract

Cloud computing is often adopted to process big data for genome analysis due to its elasticity and pay-as-you-go features. In this paper, we present SCAN, a smart application platform to facilitate parallelization of big genome analysis in clouds. With a knowledge base and an intelligent application scheduler, the SCAN enables better understanding of bio-applications' characteristics, and helps to orchestrate huge, heterogeneous tasks efficiently and cost-effectively. We conducted a simulation study and found that the SCAN platform is able to improve the performance of genome analysis and reduce its cost in a wide variety of circumstances.