No 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 fixed 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 flexibility, 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 InfiniBand, 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.

Modern applications for analysing 2D/3D data require complex visual output features which are often based on the multi-platform OpenGL® API for rendering vector graphics. Instead of providing classical workstations, the provision of powerful virtual machines (VMs) with GPU support in a scientific cloud with direct access to high performance storage is an efficient and cost effective solution. However, the automatic deployment, operation and remote access of OpenGL® API-capable VMs with professional visualization applications is a non-trivial task. In this chapter the authors demonstrate the concept of such a flexible cloud-like analysis infrastructure within the framework of the project ASTOR. The authors present an Analysis-as-a-Service (AaaS) approach based on VMware™-ESX for on demand allocation of VMs with dedicated GPU cores and up to 256 GByte RAM per machine.

No abstract

No abstract