Hi.
I have been putting off responding to the last email in the chain of our
discussions from before Christmas while I went and did a bit of
background reading. Having done this and reread the latest draft, I get
the impression that I was having a bad hair day/week or some such when I
wrote the previous review and discussed it with you. Apologies for some
rather inappropriate comments on things that I now realize were actually
pretty well covered.
Having got various other things of a gen-art nature off my job list, I
got back to AQM. I have resubmitted the review deleting the stuff about
bufferbloat, but I do think that some words about checking that
combinations of mechanisms don't mess with each other would be a good idea.
Incidentally, I thought that CableLabs report of comparative simulations
of the various proposals in train at the moment was a nice piece of
work. The background material is a useful tutorial, the analysis of the
requirements of gaming users is helpful and the comparative analysis is
very helpful. It was mentioned a couple of days ago in a thread on
IETF Discuss on a slightly related matter :
http://www.cablelabs.com/wp-content/uploads/2013/11/Active_Queue_Management_Algorithms_DOCSIS_3_0.pdf
Regards,
Elwyn
_______________________________________________
aqm mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/aqm
