Section: New Results

Energy-aware scheduling for clouds

Participants: Y. Kessaci, N. Melab, E-G. Talbi

High-performance computing (HPC) is moving from in-house to cloud-based HPC. One of the major issues of this later is the scheduling of HPC applications taking into account the energy criterion in addition to performance. In [54] , we have addressed that issue (Ph.D thesis of Y. Kessaci). We have proposed several metaheuristics for cloud managers and experimented on OpenNebula using different (VMs) arrival scenarii and different hardware infrastructures. The results show that our approaches outperform the scheduler provided in OpenNebula by a significant margin in terms of energy consumption and number of scheduled VMs.