Changes between Version 6 and Version 7 of SAW


Ignore:
Timestamp:
06/27/12 17:01:34 (12 years ago)
Author:
hdempsey@bbn.com
Comment:

add more info on VLANs and their status

Legend:

Unmodified
Added
Removed
Modified
  • SAW

    v6 v7  
    88A [http://groups.geni.net/syseng/attachment/wiki/SAW/Slice%20Around%20The%20World%20Demonstration%20Description%20V1.doc general description] is attached to this page.  It lists all the labs that intend to be part of the SAW demo in some way.  A one-sentence summary of the demo is that Starlight will engineer VLANs that span many different international research labs that have projects in network virtualization, and those projects will collaborate to run high-touch applications over dynamic processing and network resources.  The first application is supposed to be scientific visualization for distributed computing associated with nanotechnology experiments.
    99
    10 A [http://groups.geni.net/syseng/attachment/wiki/SAW/SATW-VLAN.pdf partial network diagram] showing connections between Japan, Internet2 ION, ProtoGENI Utah (only) and NREN/Starlight (show as GENI cloud) illustrates how four non-GENI VLANs (1517, 1591, 1592, 1595) will be used.  These VLANs do not overlap with any GENI mesoscale VLANs, but they do use the same ION ports that many GENI mesoscale connections use.  The initial estimated size of these VLANs was 10Gbps, but they cannot be implemented at that speed at all sites, due to shared resource constraints such as ION.
     10A [http://groups.geni.net/syseng/attachment/wiki/SAW/SATW-VLAN.pdf partial network diagram] showing connections between Japan, Internet2 ION, ProtoGENI Utah (only) and NREN/Starlight (shown as GENI cloud) illustrates how four non-GENI VLANs (1517, 1591, 1592, 1595) will be used.  Note that NREN/Starlight also goes to iCAIR, but we're not sure which iCAIR memebers are participating in the Slice Around the World demo.  These VLANs do not overlap with any GENI mesoscale VLANs, but they do use the same ION ports that many GENI mesoscale connections use.  The initial estimated size of these VLANs was 10Gbps, but they cannot be implemented at that speed at all sites, due to shared resource constraints such as ION.
     11
     12Here is some more information about VLAN numbers in Utah, Chicago and Tokyo, along with Utah test IP addresses.  Jim Chen was doing translation of the Utah VLANS, due to a previous VLAN numbering limitation.  Rob said that limitation no longer applies, but they have left it this way since it was tested and declared working on 6/25.
     13
     14Utah vlan 517, IP:10.15.17.50/24 Chicago vlan 1517, IP: 10.15.17.55/24  Tokyo vlan 1517  IP:
     15Utah vlan 591, IP:10.15.91.50/24 Chicago vlan 1591, IP: 10.15.91.56/24  Tokyo vlan 1591  IP:
     16Utah vlan 592, IP:10.15.92.50/24 Chicago vlan 1592, IP: 10.15.92.57/24  Tokyo vlan 1592  IP:
     17Utah vlan 595, IP:10.15.95.50/24 Chicago vlan 1595, IP: 10.15.95.58/24  Tokyo vlan 1595  IP:
     18
    1119
    1220A [http://groups.geni.net/syseng/attachment/wiki/SAW/NLR-Chicago-OF-iCAIR-v2.pdf partial NREN connection diagram] shows switches that may be part of SAW at NREN.  This figure was actually intended to document how the InstaGENI rack at NREN will connect to GENI--disregard the VLAN information.
     
    2836Matt Zekauskas (I2 connections, including connections to ION and ProtoGENI)
    2937
     38Unknown contact name Pacific Northwest GigaPoP (landing site for JGN connections)
     39
    3040Rob Ricci (connections to ProtoGENI and support for an NC7 VNode from University of Tokyo)
    3141