Browsing by Author "Lowe, John Michael"

Filter results by typing the first few letters
Now showing 1 - 15 of 15
  • Thumbnail Image
    Item
    Gateway Hosting at Indiana University
    (2009-06) Lowe, John Michael; Shields, Corey; Hancock, David Y.; Link, Matthew R.; Stewart, Craig A.; Pierce, Marlon
    The gateway hosting service at Indiana University provides science gateways and portals with hosting resources to facilitate the use of computation resources and storage within the TeraGrid. This service is designed with high availability in mind and is deployed across the Indianapolis and Bloomington campuses with redundant network, power, and storage. The service uses OpenVZ to give each gateway or portal its own virtual environment while making the most efficient use of the hardware and administrative resources. OpenVZ’s user beancounter quota system and fair-share scheduling for processes and I/O allows fair distribution of resource between virtual machines while allowing full utilization of the hardware. The ability to do live migration allows kernel updates without service interruption. Indiana University’s research network provides multiple low latency high bandwidth connections between campuses, other TeraGrid resource providers, and the Internet at large. The service is in use by a variety of projects such as FlyBase and TeraGrid Information Services and, since the service was put into production in August 2008, there have been 5.37 hours of down time.
  • No Thumbnail Available
    Item
    Image is Everything: Dynamic HPC VM Repositories using Murano
    (2016-10-27) Lowe, John Michael; Budden, Robert; Fischer, Jeremy
  • Thumbnail Image
    Item
    Jetstream (NSF Award 1445604) Annual Report: December 1, 2018 – November 28, 2019
    (2019-11-26) Hancock, David Y; Merchant, Nirav; Lowe, John Michael; Fischer, Jeremy; Liming, Lee; Taylor, James; Afgan, Enis; Turner, George; Skidmore, Edwin; Beck, Brain W.; Snapp-Childs, Winona; Foster, Ian; Vaughn, Matthew
  • Thumbnail Image
    Item
    Jetstream (NSF Award 1445604) Program Year 3 Annual Report (December 1, 2016 – November 28, 2017)
    (2017-12-15) Stewart, Craig A.; Hancock, David Y.; Merchant, Nirav; Lowe, John Michael; Fischer, Jeremy; Liming, Lee; Taylor, James; Afgan, Enis; Turner, George; Skidmore, Edwin; Packard, Michael; Beck, Brain W.; Foster, Ian; Vaughn, Matthew
  • Thumbnail Image
    Item
    Jetstream (NSF Award 1445604) Year Program Year 2 Annual Report (Dec 1, 2015 – Nov 30, 2016)
    (2016-12-31) Stewart, Craig A.; Hancock, David Y.; Vaughn, Matthew; Merchant, Nirav; Lowe, John Michael; Fischer, Jeremy; Liming, Lee; Taylor, James; Afgan, Enis; Turner, George; Hammond, Bret; Skidmore, Edwin; Packard, Michael; Foster, Ian
  • Thumbnail Image
    Item
    Jetstream Stakeholder Advisory Board Meeting February 2017: Presenters’ Report
    (2017-02-28) Stewart, Craig A.; Hancock, David Y.; Vaughn, Matthew; Merchant, Nirav; Lowe, John Michael; Fischer, Jeremy; Liming, Lee; Taylor, James; Turner, George; Hammond, C. Bret; Skidmore, Edwin; Packard, Michael; Miller, Therese; Foster, Ian; Rad, Paul; Mehringer, Susan
  • Thumbnail Image
    Item
    Jetstream – performance, early experiences, and early results
    (2016-07-17) Stewart, Craig A.; Hancock, David Y.; Vaughn, Matthew; Fischer, Jeremy; Cockerill, Tim; Liming, Lee; Merchant, Nirav; Miller, Therese; Lowe, John Michael; Stanzione, Daniel C.; Taylor, James; Skidmore, Edwin
    Jetstream is a first-of-a-kind system for the NSF - a distributed production cloud resource. The NSF awarded funds to create Jetstream in November 2014. Here we review the purpose for creating Jetstream, present the acceptance test results that define Jetstream’s key characteristics, describe our experiences in standing up an OpenStack-based cloud environment, and share some of the early scientific results that have been obtained by researchers and students using this system. Jetstream offers unique capability within the XSEDE-supported US national cyberinfrastructure, delivering interactive virtual machines (VMs) via the Atmosphere interface developed by the University of Arizona. As a multi-region deployment that operates as a single integrated system, Jetstream is proving effective in supporting modes and disciplines of research traditionally underrepresented on larger XSEDE-supported clusters and supercomputers. Already, researchers in biology, network science, economics, earth science, and computer science have used Jetstream to perform research – much of it research in the “long tail of science.”
  • No Thumbnail Available
    Item
  • No Thumbnail Available
    Item
    Jetstream: A science and engineering cloud
    (2016-11-16) Lowe, John Michael
  • No Thumbnail Available
    Item
    NSF High Performance Computing System Acquisition system description: Jetstream - a self-provisioned, scalable science and engineering cloud environment
    Stewart, Craig A.; Stanzione, Daniel; Cockerill, Timothy; Skidmore, Edwin; Fischer, Jeremy; Lowe, John Michael; Hammond, Bret; Turner, George; Hancock, David Y.; Miller, Therese
  • No Thumbnail Available
    Item
    OpenStack deployment for Jetstream
    (2016-11-15) Lowe, John Michael
  • Thumbnail Image
    Item
    Performance Characteristics of Virtualized GPUs for Deep Learning
    (2019-10) Michael, Scott; Teige, Scott; Li, Junjie; Lowe, John Michael; Turner, George; Henschel, Robert
    As deep learning techniques and algorithms become more and more common in scientific workflows, HPC centers are grappling with how best to provide GPU resources and support deep learning workloads. One novel method of deployment is to virtualize GPU resources allowing for multiple VM instances to have logically distinct virtual GPUs (vGPUs) on a shared physical GPU. However, there are many operational and performance implications to consider before deploying a vGPU service in an HPC center. In this paper, we investigate the performance characteristics of vGPUs for both traditional HPC workloads and for deep learning training and inference workloads. Using NVIDIA’s vDWS virtualization software, we perform a series of HPC and deep learning benchmarks on both non-virtualized (bare metal) and vGPUs of various sizes and configurations. We report on several of the challenges we discovered in deploying and operating a variety of virtualized instance sizes and configurations. We find that the overhead of virtualization on HPC workloads is generally < 10%, and can vary considerably for deep learning, depending on the task.
  • Thumbnail Image
    Item
    Resource Management from HPC to the Cloud: Do you manage resources or do they manage you?
    (2016-06-14) Hancock, David Y.; Stewart, Craig A.; Fischer, Jeremy; Lowe, John Michael; Rad, Paul; Vaughn, Matthew
  • Thumbnail Image
    Item
    Security Best Practices for Academic Cloud Service Providers
    (2018-05) Dooley, Rion; Edmonds, Andy; Hancock, David Y.; Lowe, John Michael; Skidmore, Edwin; Adams, Andrew K.; Kiser, Ryan; Krenz, Mark; Welch, Von; Knepper, Richard
  • Thumbnail Image
    Item
    System Acceptance Report for NSF award 1445604 ”High Performance Computing System Acquisition: Jetstream - A Self-Provisioned, Scalable Science and Engineering Cloud Environment”
    (2016-05-11) Stewart, Craig A.; Hancock, David Y.; Vaughn, Matthew; Merchant, Nirav C.; Lowe, John Michael; Fischer, Jeremy; Liming, Lee; Taylor, James; Afgan, Enis; Turner, George; Hammond, C. Bret; Skidmore, Edwin; Packard, Michael; Foster, Ian