Cluster D conference call on Friday, June 18, 2010


Attendees:   [[BR]]

Ilia Baldine  (RENCI)[[BR]]
Anirban Mandal [[BR]]
Chris Heermann[[BR]]

Harry Mussman  (GPO)  [[BR]]



Mike Zink (UMass) [[BR]]
Emmanuel Cecchet [[BR]]
Brian Lynn[[BR]]
David Irwin[[BR]]

Denise Gurkan  (Univ Houston)

Michael Wang  (Columbia) [[BR]]

Onjin  (Oklahoma)  [[BR]]
Wenjie[[BR]]

Wenjie (Ohio State)  [[BR]]

Hongwei	(Wayne State)[[BR]]

Rudra Dutta  (NCState) [[BR]]

Agenda:  [[BR]]

1) Taking stock of known unresolved problems:[[BR]]
+  Brian Lynn:  Still has issues with pre-mature close[[BR]]
+  Ilia:  Study more.[[BR]]
+  David:  Eucalyptus cluster at UMass;  need to connect via new clearinghouse, before GEC8 [[BR]]
+  Ilia:  Have clearinghouse up by next week. [[BR]]
+  David:  Needs to check new broker, explain issues with policy and make request. [[BR]]

2) New networking driver structure[[BR]]
+  Denise:  Shade working on driver for 3400 [[BR]]
+  Ilia:  New document on drivers, quite detailed. [[BR]]

3) GEC8 demos  [[BR]]
+  DICLOUD:  yes;  ticket [[BR]]
+  ViSE:  yes;  no ticket [[BR]]
+  DOME:  no [[BR]]
+  IMF:  yes [[BR]]
+  ORCA/BEN: [[BR]]
-  OpenFlow demo RENCI and GIST Korea [[BR]]
-  Programmable router, on multiple site hosts,  practice for GEC9 [[BR]]
+  Kansei:  yes;  un cross-site demo, sensor nodes and latency through  [[BR]]
+  OKGEM:  yes, possibly controllable robot [[BR]]
+  LEARN:  Integration of 3400 into LEARN, at Univ Houston and Rice [[BR]]

4)  GENI Solictation 3:  [[BR]]
+  Ilia:  Considering GENI racks [[BR]]
-  Many Cluster D sites [[BR]]
-  Each site with separate network connectivity   [[BR]]
-  Spaces in lab or university data center?   [[BR]]
-  Connectivity via many (10+) VLANs?   [[BR]]
-  Need good connectivity into sites   [[BR]] 
+  David:  Space, power, cooling requirements for each rack? [[BR]]
-  Ilia:  19" full height rack [[BR]]
-  GENI enabled open-flow switches [[BR]]
+  Ilia:  Purchased by project, sent out to each site [[BR]]
-  Agreement with manufacturer, who sends it out. [[BR]]
-  Harry:  See current WiMAX mesoscale project  [[BR]]
+  Ilia:  Two CFs:  ORCA and protoGENI  Can be accessed by either [[BR]]
-  Local interface, but also managed by central site (e.g., RENCI) [[BR]]
+  Ilia:  How many sites?  40 racks too many for level of funding [[BR]]
-  Perhaps 15 sites, each with one or two racks [[BR]]
-  Need sites on the west coast [[BR]]
+  Ilia:  Connectivity to backbone costs [[BR]]
-  Campus?    [[BR]]
-  Regional networks?     [[BR]]
-  Perhaps:  FrameNet wherever possible, otherwise Starlight [[BR]]


Last call A) Connectivity to Wayne State and Ohio State[[BR]]
+  Joe:  Ready to go from Starlight to Merit racks; [[BR]] 
+  Hongwei:  Merit to OARnet to Ohio State;  also Merit to Wayne State[[BR]]
+ Ilia:  Working with MAX on interoperability[[BR]]
+ Ilia:  working toward translation from protoGENI to NDL  [[BR]]
- Harry: Srikanth also working on topic[[BR]]

Last call B)  GEC9 demo proposals, accepted: [[BR]]
a)  Joe (Northwestern):  iGENI, HP and Ericsson:  video transcoding on clouds over dynamically provisioned networks.  Prototype at GEC8.  HP wants to include OpenFlow. [[BR]]
b)  David (UMass):  Now-casting (short-term weather forcasting), GEC7 demo plus more computation [[BR]]
c)  Jeff Chase (Duke):  Dynamically stand-up ASs with custom click routers	[[BR]]