Fri: I'd like to spend some time on Friday constructing
a list of "little" studies we could do. We've already
identified a few, eg, the measurement of the utility
of detour routing, the BGP study.
I am particularly interested in coming up with simulation
studies we (well, ok, Becker) could do, eg, can we demonstrate that a
large number of short connections causes congestion problems
that can only be solved by traffic shaping? Or can we show
that TCP performance is always improved by giving acks priority?
What I'd like us all to do is to spend some time thinking
about this, and come to Friday's meeting with a list of
suggestions. We could then brainstorm how we would go about
conducting each study.
tom