A feasibility study of using HPC cloud environment for seismic exploration

Abstract

Seismic applications require High Performance Computing (HPC) to simulate the dynamics of complex models. Not a long ago, companies, research institutes, and universities used to acquire clusters of computers to maintain them on-premise. Recently, cloud computing has become an alternative for a subset of HPC applications, including seismic. This poses a challenge to the end-users, who have to decide whether they should run their applications on their local clusters or burst them to a remote cloud provider. In this paper, we present a feasibility study of using cloud for seismic processing, comparing its performance with a default on-premise HPC cluster. For that, an architecture for FWI as a service on cloud is proposed and the paper discusses also the possible financial benefits of cloud. Results indicate that cloud can be an interesting alternative to boost on-premise HPC clusters. It achieves similar or even lower turnaround time in comparison to HPC clusters, avoiding queue waiting time.