Rong,
I promised my proof that code to de-randomize drops could be pointless.
It's within S.7.7.1 of my PhD dissertation:
<http://www.bobbriscoe.net/pubs.html#refb-dis>
Search for the heading:
"Are RED Markings Uniformly Distributed"
Admittedly, the proof depends on the assumption "as long as the flow
rate is small relative to the link capacity", which might not be
applicable in your deployment scenario. However, it might be worth
you taking a look.
Bob
________________________________________________________________
Bob Briscoe, BT
_______________________________________________
aqm mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/aqm
