Paris-Networking

About Paris-Networking | Announce a talk | Subscribe

Seminar: Two Problems in Network Probing  

Darryl Veitch, The University of Melbourne

Wednesday, June 27th 2007, 11h00 - 12h00

Location :

Thomson Paris Lab
46, Quai A. Le Gallo
92648 Boulogne Cedex
Nearest metro station: "Boulogne Pont de Saint Cloud" (line 10).
Look for  directions  with mappy.fr.

Please register in advance!
All are welcome to attend the sessions and stay for lunch, but we would like to know in advance to help organizing the event.
Please contact letizia.ballarini@thomson.net or augustin.chaintreau@thomson.net, to announce that you are coming. 
Thank you.

Abstract :

Two Problems in Network Probing

Abstract:
In this talk I will report on two pieces of recent work in the field of probing
based network measurement.

The first is a topic in multicast based loss tomography. We show for the first
time how temporal properties of link loss processes can be inferred, based on
observed losses at the receiver leaves of a multicast tree.  The second
concerns optimal measurement of loss and delay based on probing.  We show how,
provided the underlying process being measured obeys a convexity property, that
the optimal probing stream with respect to the sampling problem is 'universal'
rather than being a function of the system being probed, and show how the
optimal result can be approached in practice.

Host :

François Baccelli and Christophe Diot