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

[Sally Floyd]

>And you forgot about some of SQ's benefits:
>
>  SQ only requires a single message on the uncongested leg, rather
>  than a round-trip through the known congested leg.

Why do you assume that the path from the congested router back to
the source is a single leg that is itself uncongested?  The congested
router setting the ECN bit or sending the source quench could perhaps
be on the other side of the globe from the source, sitting at the
last hop to the receiver, with the reverse path from that router
back to the source containing several congestion points (the
transoceanic link, the public exchange point, and the last hop to
the sender itself) where a packet is using scarce bandwidth, and
could itself be dropped.  At the very least, this possibility needs
to be taken into account.

- Sally
--------------------------------
http://www.aciri.org/floyd/
--------------------------------