[e2e] Can feedback be generated more fast in ECN?

[Jonathan Stone]

In message <200102212103.f1LL3PA07174 at calcite.rhyolite.com>,
Vernon Schryver writes:

>> From: "Eric A. Hall" <ehall at ehsco.com>

[...]

>Some of what both SQ and ECN can do is clear.  We agree that SQ is
>somewhere beteen much faster and the same speed of ECN, depending on the
>location of the bottleneck, with a likely case being no speed difference
>because the bottleneck is at the far edge of the net.  SQ pays for that
>possible speed advantage by spending bandwidth and router CPU cycles, as
>well as possible new security holes.

Is the edge-of-the-net case really that likely? Modern, New-Reno
vintage TCPs seem to be fairly good at converging on close to fair
shares, even on wet-string links like cellphones.

Congestion is known to happen at some relatively ``close'' points,
like cross-ocean links and campus border gateways flooded by Napster
traffic.

Is there any hard data on the distribution of congestion actually
occurs-- i.e., just where either SQ or ECN would come into effect?
Off-hand, it seems hard to gather good data without using a mechanism
which (like SQ) tells you the address of the congestion-detecting point.