Sept. 29, 2006

From ReddNet
Jump to navigation Jump to search

Probable Attendees

  • Blackwell (SFA)
  • Hagewood, Assad (Nevoa)
  • Beck, Moore(UTK)
  • Sheldon, Tackett, Pathak, Brown (VU)
  • Swany (UDel)

Current Agenda (Start time this week, 10:30EDT/9:30CDT)

Application Communities

  • AmericaView (PR)

Had several conversations with the board and some of the members.

Deployment and Operations

  • Current status of deployment effort: Fermi, ORNL, etc. (Bobby)
  • Do we have a chance of getting a REDDnet node at CERN (Paul, Martin)
  • Status of AMPATH link(Surya)
    • Had a general discussion of the problem of long RTT and latency on this link.
    • We need to get better data. But the key point is that we don't seem to be making any progress in this area. We seem to be blocked by the "dirty little secret" of HP networking. The approach that depends on careful control, expert tuning of the backbone routers won't scale, and we may have encountered that here.
    • Current solutions seem to depend on artificial tricks that won't scale.
    • Plan: Set it up to run at whatever rate we can get (400 Mbps) and let it run. Utilization is what people are looking for, so let's do that.
  • After SC, all of the hardware is going to be shipped back to Vandy.

Technology: Hardware, Software

  • Current state of the components of the L-Store software Suite(Alan, et. al)
  • Availability of alternative versions of the IBP depot code. (Micah)
  • Current state of the Nevoa components and tools (Hunter)
  • Agreement to discuss the formulation of a standard directory protocol in at the Chicago meeting (i.e. on Dec. 3, the day before the large meeting).

Organization and Funding Opportunities

  • Status of NSF PIRE
    • Vandy has given approval to go ahead on the NSF PIRE
    • 6 page preprop due in Oct. 30.
  • What's happening with the SBIR?
    • Envelock systems is working on back-up capabilities under L-Store IBP
    • Due Date: December 4

Events, Education, Outreach

  • SC06?
    • An SC06 page has been added to the Wiki (linked from "REDDnet at work")
    • Plans for hardware and demonstrations. Need a good description of what will be demonstrated as the Vanderbilt booth so that we can distribute it to people we want to come see what we can do.
    • Description of Jian's Demo:

Creating a multiuser visualization facility based on distributed and heterogeneous resources has proved to be an elusive challenge for the high performance computing and visualization community. We have developed a highly deployable solution to this problem. Our solution uses a collection of freely available, unscheduled and unreserved storage and computing resources, distributed across the network, to provide the essential infrastructure that such a visualization facility needs. Our demonstration will show that large scale visualization of cutting edge simulation datasets can be shared among multiple distributed users from the same pool of distributed resources, requiring only a standard Internet connection for each user.

    • Working with Ultralight? We don't know what's being planned here.
    • Working with OSG/Fermi? We don't know what they're planning.
    • Martin is doing a bandwidth challenge with Phoebus. It's not all-TCP-Phoebus. This demo is with Internet2.
    • Engaging with ORNL about standing up an L-Store/SRM at their site.
    • Status of Ultralight plans (Alan)
  • Plans for Internet2 Side meeting (Terry, Micah)
    • Hunter has talked to the RNP people and they're interested in attending, but details aren't clear at this point.

Action Items