[e2e] Open the floodgate

[Ted Faber]

On Sat, Apr 24, 2004 at 02:51:12AM -0700, Sam Manthorpe wrote:
> This is probably a dumb question, since I've been out of networking
> for quite a while, but doesn't the TCP/IP stack *have* a network-layer
> congestion amerlioration mechanism: ICMP source quench. 

In addition to your concerns about how an end host should respond to an
ICMP source quench, sending more packets during congestion is not
exactly fail safe.  If getting the quench means slow down, not hearing
one is implicitly an encouragement if not a request to speed up.
Encouragements to speed up in the face of loss can lead to congestion
collapse if that loss is due to congestion.  Systems that (at least)
don't speed up in the absence of messages from the network tend to be
more fail safe.

-- 
Ted Faber
http://www.isi.edu/~faber           PGP: http://www.isi.edu/~faber/pubkeys.asc
Unexpected attachment on this mail? See http://www.isi.edu/~faber/FAQ.html#SIG 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://www.postel.org/pipermail/end2end-interest/attachments/20040426/523d369d/attachment.bin