Re: Opinions regarding Packeteer Packetshaper & alternatives

[Mark Poepping <poepping () CMU EDU>]

I imagine it's been a while since this information was circulated, but for
those considering hardware bandwidth management appliances such as the
packeteer and netenforcer, I refer you to at least one possible alternative,
an approach we discussed several years ago at JointTechs: social engineering
(it worked as well for us as it does for the bad guys:-).

The presentation materials are at:  http://www.net.cmu.edu/pres/jt0803/

The site also has our technical comparison of the devices available at the
time, we bought a netenforcer and didn't mind it, but engaging the users
turned out to be *way* more effective, and we really haven't had a bandwidth
problem since.  The approach essentially requires that you "tell users"
how much bandwidth they're actually using. It requires that you count and
account for the usage (we use argus at the egress, but netflow works too - I
think Cornell uses netflow for similar purposes, though their method
includes actually 'charging'), and as long as you have a way to attribute IP
traffic to 'users', you're good to go.  The software to implement our
mechanism isn't quite available (we just didn't push to publish it), but
it's part of the CMU netreg suite (predates the 'other' netreg)..

I wish more problems were this 'easy'..
Mark.