Incidentally I found this paper in ACM SIGMETRICS 1992 covering more or less
precisely the same algorithm we're using for our clock sweep. I haven't quite
digested it yet myself so I'm not sure what the conclusions about weights tell
us to do with our buffer usage counter.
I put a copy up for download since even though it's permitted to copy I don't
know how to find a public link. It's kind of big so please download it and
read it locally, don't try to read it from my machine:
http://stark.xeocode.com/~stark/p35-nicola.pdf.gz
Regarding copyright the paper sez:
Permission to copy without fee all or part of this material is
granted provided that the copies are not made or distributed for
direct commercial advantage, the ACM copyright notice and the
title of the publication and its date appear, and notice is given
that copying is by permission of ths Association for Computing
Machinery. To copy otherwise, or to republish, requires a fee
and/or specific permission.
-- Gregory Stark EnterpriseDB http://www.enterprisedb.com