Hyper Link:
IGI Global

Abstract:
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.

BibTeX Code:
@article{mexner2017opengl,
title={OpenGL{\textregistered} API-Based Analysis of Large Datasets in a Cloud Environment},
author={Mexner, Wolfgang and Bonn, Matthias and Kopmann, Andreas and Mauch, Viktor and Ressmann, Doris and Chilingaryan, Suren A and Jerome, Nicholas Tan and van de Kamp, Thomas and Heuveline, Vincent and L{\"o}sel, Philipp and others},
journal={Design and Use of Virtualization Technology in Cloud Computing},
pages={161},
year={2017},
publisher={IGI Global}
} 

ISBN: 978-1-61208-388-9, Pages 66-69

Hyper Link:
ResearchGate, Paper Award

Abstract:
Today, most high performance computing (HPC) systems are equipped with high-speed interconnects providing low communication and synchronization latencies in order to run tightly coupled parallel computing jobs. They are typically managed and operated by individual institutions and offer a fixed capacity and static runtime environment with a limited selection of applications, libraries and system software components. On the contrary, a cloud-based Infrastructure-as-a-Service (IaaS) model for HPC resources promises more flexibility, as it enables elastic on-demand provisioning of virtual clusters and allows users to modify the runtime environment down to the operating system level. The goal of this research effort is the general demonstration of a prototypic HPC IaaS system allowing automated provisioning of virtualized HPC resources while retaining high and predictable performance. We present an approach to use high-speed cluster interconnects like InfiniBand within an IaaS environment. Our prototypic system is based on the cloud computing framework Openstack in combination with the Single Root - I/O Virtualization (SR-IOV) mechanism for PCI device virtualization. Our evaluation shows that, with this approach, we can successfully provide dynamically isolated partitions consisting of multiple virtual machines connected over virtualized InfiniBand devices. Users are put in the position to request their own virtualized HPC cluster on demand. They are able to extend or shrink the assigned infrastructure and to change the runtime environment according to their needs. To ensure the suitability for HPC applications, we evaluate the performance of a virtualized cluster compared to a physical environment by running latency and High-Performance Linpack (HPL) benchmarks.

BibTeX Code:
@article{mauch2015deployment,
title={Deployment of Virtual InfiniBand Clusters with Multi-tenancy for Cloud Computing},
author={Mauch, Viktor},
journal={CLOUD COMPUTING 2015},
pages={81},
year={2015}
} 

Hyper Link:
PCaPAC 2014, Session: FC02 - User Interfaces

Abstract:
Modern data analysis applications for 2D/3D data samples require complex visual output features which are often based on OpenGL, a multi-platform API for rendering vector graphics. They demand special computing workstations with a corresponding CPU and GPU power, enough main memory and fast network interconnects for a performant remote data access. For this reason, users depend heavily on available free workstations, both temporally and locally. The provision of virtual machines (VMs) accessible via a remote connection could avoid this inflexibility. However, the automatic deployment, operation and remote access of OpenGL-capable VMs with professional visualization applications is a non-trivial task. In this paper, we discuss a concept for a flexible analysis infrastructure that will be part in the project ASTOR, which is the abbreviation for “Arthropod Structure revealed by ultra-fast Tomography and Online Reconstruction”. We present an Analysis-as-a-Service (AaaS) approach based on the on-demand allocation of VMs with dedicated GPU cores and a corresponding analysis environment to provide a cloud-like analysis service for scientific users. 

BibTeX Code:
@article{mauch2014opengl,
title={OpenGL-BASED data analysis in virtualized self-service environments},
author={Mauch, Viktor and Bonn, M and Chilingaryan, S and Kopmann, A and Mexner, W and Ressmann, D},
journal={Proc. PCaPAC2014, http://jacow. org},
year={2014}
}

EAN: 4038858091822, Page 96-98

Hyper Link:
heise shop

Abstract:
Neben einer Reihe von Softwareprodukten zum Aufsetzen eines Cloud-Systems hat sich OpenStack etabliert. Diese Position verdankt das Open-Source-Werkzeug vor allem einem starken Engagement nahmhafter Organisationen und Firmen. Eine Analyse deckt die Struktur und Optionen auf.

BibTeX Code: 
@article{96--98|iX special 3/2012,
author = {Tobias Kurze, Viktor Mauch},
title = {Fahrt aufnehmen},
subtitle = {OpenStack – Open-Source-Software zur Cloud-Steuerung},
journal = {iX special},
volume = {3},
year = {2012},
pages = {96--98},
} 

ISBN: 978-1-4503-1161-8, Article No. 9

Hyper Link:
ACM Digital Library

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.

BibTeX Code:
@inproceedings{hillenbrand2012virtual,
title={Virtual InfiniBand Clusters for HPC clouds},
author={Hillenbrand, M. and Mauch, V. and Stoess, J. and Miller, K. and Bellosa, F.},
booktitle={Proceedings of the 2nd International Workshop on Cloud Computing Platforms},
pages={9},
year={2012},
organization={ACM}
} 

ISSN: 0935-9680, Page 72-73

Hyper Link:
heise online

Abstract:
Neben den kommerziellen Verwaltungshilfsmitteln für die Cloud haben sich inzwischen Open-Source-Werkzeuge etabliert. Großen Zuspruch fand OpenNebula, das jetzt in einer neuen Version vorliegt.

BibTeX Code:
@article{72--73|iX 4/2012,
author = {Viktor Mauch, Marcel Kunze},
title = {Freie Fahrt},
subtitle = {Upgrade der Cloud-Management-Plattform OpenNebula},
journal = {iX},
volume = {4},
year = {2012},
pages = {72--73},
}