Sunday, January 16, 10:00-11:00
at 10:00
Schreiber Building, Room 309
In this talk I will describe short- and long-term research contributions
to QoS routing. I will start with a practical QoS-aware multicast
routing
protocol that can be deployed in today's Internet. The protocol
is superior
to previous suggestions in its capability to achieve high success
ratio
at close to optimal overhead, and in its ability to self-tune according
to the
tightness of the demand. Then I'll describe approximations
for
optimal QoS routing and partition for additive requirements (such
as delay).
I'll present Fully Polynomial Approximation schemes for these problems
for both the unicast and multicast cases. I'll concentrate
on the case
where the delay-cost functions are discrete, which is the most
significant in addressing the future needs of Internet QoS routing.
For colloquium schedule, see http://www.math.tau.ac.il/~zwick/colloq.html