It seems to me that a lot of what causes working group lists to
melt down is simply the volume of traffic -- usually with plenty
of off-topic banter, or exchanges of dubious value, with the resulting
conjestive collapse of our wetware buffering. On good days, the
drop algorithm may be more sophisticated than tail drops; on
bad days...
Perhaps we should take a lesson from TCP and set a receive window
on IETF mailing lists in the face of conjestion. The sender is thus
obligated to keep the transmission within the window, and as a side
effect to consider the quality of the, um, quantity. Just this simple
step would greatly limit (purposeful) DOS attacks and other death
spirals. It also mitigates the "free speech" attacks by not throttling
based on content (which is inherently contentious), but based on
wg mailing list "bandwidth".
in all modesty, Mike
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf