Data-Intensive HPC Tasks Scheduling with SDN to Enable HPC-as-a-Service
Advances in Cloud Computing attracted scientists to deploy their HPC applications to the cloud to benefit from the flexibility of the platform such as scalability and on-demand services. Nevertheless, HPC applications can face serious challenges on the cloud that could undermine the gained benefit, if care is not taken. This paper starts first comparing the performance of several HPC benchmarks on a commodity cluster and Amazon public cloud to illustrate the confronted challenges. To mitigate the problem, we have introduced a novel approach called ASETS, "A SDN Empowered Task Scheduling System", to schedule data-intensive High Performance Computing (HPC) tasks on a Cloud environment. In this paper, we focus on the implementation and performance analysis of ASETS and its first algorithm called SETSA, (SDN Empowered Task Scheduling Algorithm). ASETS uses the "bandwidth awareness" capability of SDN to better utilize network bandwidths when assigning data intensive tasks to virtual machines (workers) in the cloud. This novel approach aims to improve the performance of HPC applications on the cloud in order the platform could provide efficient HPC-as-a-Service (HPCaaS). The paper briefly describes the ASETS architecture and its SETSA algorithm, and then focuses on the details of the implementation and performance analysis of ASETS and SETSA. Preliminary results indicate that ASETS provides substantial performance improvement for HPCaaS as the degree of multi-tenancy in the cloud increases. This result is significant since it indicate both the users and the cloud service providers can benefit from ASETS.