Bandwidth-Aware Resource Management for Extreme Scale Systems

As systems scale towards exascale, many resources will become increasingly constrained. While some of these resources have historically been explicitly allocated, many, like network bandwidth, I/O bandwidth, or power, have not. As systems continue to evolve, we expect many such resources to become explicitly managed. This change will pose critical challenges to resource management and job scheduling. In this paper, we explore bandwidth-aware resource management for Blue Gene systems, where the partition-based interconnect architecture provides a unique opportunity to explicitly allocate bandwidth to jobs. In this paper we investigate the value of bandwidth awareness and further present a bandwidth-aware resource management design for Blue Gene systems.