Fault-tolerant and reliable computation in cloud computing

Abstract

Cloud computing, with its great potentials in low cost and on-demand services, is a promising computing platform for both commercial and non-commercial computation clients. In this work, we investigate the security perspective of scientific computation in cloud computing. We investigate a cloud selection strategy to decompose the matrix multiplication problem into several tasks which will be submitted to different clouds. In particular, we propose techniques to improve the fault-tolerance and reliability of a rather general scientific computation: matrix multiplication. Through our techniques, we demonstrate that fault-tolerance and reliability against faulty and even malicious clouds in cloud computing can be achieved.