linearly and packet drops as 1/sqrt(p) (ie chose a shorter path over an
equally smaller packet drop rate). In truth though, this is a flawed
metric without a model of the resource usage on the link. Both the
latency and the drop rate are dependent on the amount of flow pushed
down the path. We don't know what that function is. My intuition is
that the optimal tradeoff between RTT and p will change over their
range. My GUESS is that for low values of p, reductions in RTT will
matter much more, and for high values of p reductions in p will count
more. I expect this effect will be much stronger for short flows than
long.
However, as Neal wrote, I suggest using the the linear p, inverse
quadratic p metric for now... and lets see what the graph looks like.
- Stefan
-----Original Message-----
From: John Snell [mailto:geigudr@cs.washington.edu]
Sent: Tuesday, October 06, 1998 10:41 AM
To: syn@cs.washington.edu
Subject: best analysis metric, next measurement phase?
Question:
The next phase of the measurement study has me generating a runtime
graph
of the network, whose edges are weighted acording to some measured
property -- ie, latency, loss, hops, etc.
Previously I'd thought to separately use both latency and loss (ie, do
multiple alternate path measurements for the same path). This is
proving
to be too expensive, pushing me towards using one single metric.
Tom had recommended using the Savage/Cardwell power metric. Anyone else
have suggestions on a good metric?
-John
---------- Forwarded message ----------
Date: Mon, 5 Oct 1998 15:51:18 -0700 (PDT)
From: Tom Anderson <tom@emigrant>
To: geigudr@cs.washington.edu
Subject: Re: one other thing...
i got hammered at MIT for the fact that the latency graph
is optimized for latency, and the drop rate graph is optimized
for drops, but there is no guarantee you can get both.
I'd rather get group feedback on this one. Can you pose it as
a design question to the group?
tom