July 29, 2010

Attendees: Dantong, Dimitrios, Xin, Junmin, Viji, Alex

Testbed

We do not have the testbed at this moment because Internet2 changed their network to Juniper, from DCN to ION.
The peering with ESnet was restored. The Merit network (Michigan regional network) does not have the peering point yet. They plan to fix it in a week (8/5/2010).
We are getting some new nodes at University of Michgan (two 10Gbps nodes ). We have to set up the GridFTP servers again.
We are waiting for a new node (new control node) at BNL.
We plan to bring up a secondary terapaths service node(testbed of the testbed).
The maximum bandwidth with the new nodes is 10Gbps. With the current node, we can do 1 Gpbs. We start small, and ramp up afterward.

three cases: one request with one file, one request with multiple files, and multiple requests each having multiple files.

For test case 2.3. Will the different users be treated same or different?
(Dimitrios) The user management part in TeraPaths is under development. At this moment, the different clients are treated in the same way.

Some discussion about the ephemeral port range and how to differentiate different requests which might use the same ephemeral port.

BestMan client part is ready. Internal change is completed. BeStMan is able to talk to the TeraPaths API.

Dantong will arrange with the BNL site coordinator for the demo at BNL booth.
source and destination between BNL and University of Michigan.
We need to show how the reservation is going; the status of each fragment of the network.
We also show how BeStMan uses the reserved bandwidth.

Possibility of using Dan Gunter's NetLogger server based on GridFTP logs - showing web based graphs.
A web based graphical interface is going to be created to display the ganglia plots.