Here is my running summary of this debate, with only mild editorizing:
From: alan@mindvision.com (Alan Hannan)
I believe DIGEX should be congratulated on their forward exploration
of technology. Their creative solutions to problems like this are the
spirit that will sustain the Internet.
DWDM and huge fiber builds will get us part of the way to handling
internet growth, but we need to find ways to utilize upper-layer
tweaks to do more with less.
Could we get the summary and insight at the beginning of the note next
time? 
Thanks!
It's worth noting that as pipes get bigger and faster, caching
assumes a greater, not lesser, importance. Higher bandwidth
and constant latency means a greater bandwidth-delay product,
and a corresponding degradation in performance. Caching is
just as much about local replication to reduce latency as it is
about "conserving bandwidth."
Peter Berger
Coordinator, NLANR Engineering Services
NCNE / National Center for Network Engineering
Pittsburgh Supercomputing Center
It's worth noting that as pipes get bigger and faster, caching
assumes a greater, not lesser, importance. Higher bandwidth
and constant latency means a greater bandwidth-delay product,
and a corresponding degradation in performance. Caching is
just as much about local replication to reduce latency as it is
about "conserving bandwidth."
Particularly when a large proportion of the stacks out there perform
very poorly under the prevailing conditions in today's Internet with
its medium-to-high, jittery latencies and occasional to frequent
packet loss (cough, cough, Redmond, cough). Far better to use a proxy
on a platform like BSD that at least doesn't have an egg-sucking TCP
implementation.
---Rob