System-Level Virtualization for High Performance Computing

Abstract

System-level virtualization has been a research topic since the 70's but regained popularity during the past few years because of the availability of efficient solution such as Xen and the implementation of hardware support in commodity processors (e.g. Intel-VT, AMD-V).However, a majority of system-level virtualization projects is guided by the server consolidation market. As a result, current virtualization solutions appear to not be suitable for high performance computing (HPC) which is typically based on large-scale systems. On another hand there is significant interest in exploiting virtual machines (VMs) within HPC for a number of other reasons. By virtualizing the machine, one is able to run a variety of operating systems and environments as needed by the applications. Virtualization allows users to isolate workloads, improving security and reliability. It is also possible to support nonnative environments and/or legacy operating environments through virtualization. In addition, it is possible to balance work loads, use migration techniques to relocate applications from failing machines, and isolate fault systems for repair.This document presents the challenges for the implementation of a system-level virtualization solution for HPC. It also presents a brief survey of the different approaches and

Towards high-available and energy-efficient virtual computing environments in the cloud

“…(4) defines the minimum resources r(t), in Mflops/s, assigned to a task t that are necessary to complete its remaining workload W (t, µ) in the time from time µ + m ei to the task deadline d t . For example, consider the matrix multiplication algorithm that requires 2n 3 flops for matrices of size (n, n) to compute. If the amount of time available to execute the task is 120 s, the required resource is 2n 3 /120 flops/s.…”

“…For example, consider the matrix multiplication algorithm that requires 2n 3 flops for matrices of size (n, n) to compute. If the amount of time available to execute the task is 120 s, the required resource is 2n 3 /120 flops/s. Likewise Stillwell et al [29], we assume that a task t only requires the maximum amount of resources max r(t) necessary to execute at the maximum speed, a value defined by the user.…”

Improving Service Performance in Cloud Computing with Network Memory Virtualization

“…Interestingly, the re-emergence of virtualization [36,37] with its usage in the High Performance Computing (HPC) and grid environment show a great potential as a method for enhancing cloud computing performance. Vallee et.…”

Design and implementation of a cloud computing service for finite element analysis

“…The service is currently not available. Recent research also suggests that IaaS-based HPC cloud services [31,32] can suffer in performance especially due to VM resource competition and lack of low-latency interconnection needed by specialized parallel engineering simulations.…”