Abstract:

High Performance Computing (HPC) employs fast interconnect technologies to provide low communication and synchronization latencies for tightly coupled parallel compute jobs. Contemporary HPC clusters have a xed capacity and static runtime environments; they cannot elastically adapt to dynamic workloads, and provide a limited selection of applications, libraries, and system software. In contrast, a cloud model for HPC clusters promises more exibility, as it provides elastic virtual clusters to be available on-demand. This is not possible with physically owned clusters.

In this paper, we present an approach that makes it possible to use InfiniBand clusters for HPC cloud computing. We propose a performance-driven design of an HPC IaaS layer for In niBand, which provides throughput and latency-aware virtualization of nodes, networks, and network topologies, as well as an approach to an HPC-aware, multi-tenant cloud management system for elastic virtualized HPC compute clusters.

BibTex:

@inproceedings{hillenbrand12hpcclouds,
   author = {Marius Hillenbrand, Viktor Mauch, Jan Stoess, Konrad Miller and Frank Bellosa},
   title = {Virtual InfiniBand Clusters for HPC Clouds},
   booktitle = {Proceedings of the 2nd International Workshop on Cloud Computing Platforms (CloudCP 2012)},
   address = {Bern, Switzerland},
   year = 2012,
   month = apr # "~10",
   affiliation = {Karlsruhe Institute of Technology, Germany},
 }