Is anyone actually USING IP QoS?

Sean Donelan SEAN at SDG.DRA.COM
Wed May 19 10:16:34 UTC 1999

nbastin at rbbsystems.COM (Nick Bastin) writes:
>I'll concede this point.  However, it's worth pointing out that
>megabytes of disk space can actually be 'had', while bandwidth doesn't
>really exist.  Most of us, at least, have to pay for bandwidth by the
>month, while we only pay for the same piece of storage space once (and
>then we actually posess it).  This is important because it impacts how
>far this cost model can be extended, and for how long.

Once upon a time you used to buy CPU seconds from the campus computer
center at an exhorbitant rate.  Now you just buy the whole CPU, and waste
most of it running a screen saver of South Park characters making rude

Actually you can buy "bandwidth" much the same way.  Stop thinking
about buying 'bandwidth' in T1 or E3 size chunks, and instead think about
buying 15,000 route miles of fiber for the flat fee of $20million.  Ok,
so that seems like a lot of money.  Think of the difference between
renting an apartment or buying a house, structure the financing correctly
and the monthly payments are fairly reasonable.  Throw in DWM advances
in the last couple of years, and you can now sell 80 times more.
Trans-oceanic fiber costs more, about $150million, and more political
hurdles, but you can do much the same thing.

The 'holy grail' of QOS has little to do with delivering bits, and a lot
to do with pricing models.  Personally I think QOS could be very successfull
at the customer ingress/egress points, but it doesn't make a whole lot of
sense anywhere in the middle.  QOS is all about letting the customer buy
more or less service than they could normally afford.   Like the used
car salesperson, the carrier wants to find the largest easy monthly payment
you are willing to pay, for the worst piece of junk he can move off his
lot.  But in the middle, it is almost always cheaper and more efficient
for the carrier just to throw more bandwidth at the problem.  Bandwidth
becomes a sunk cost, with a very short shelf-life.

Next week, most of the QOS folks will be meeting down in San Francisco
instead of Eugene.

QOS has become such a marketing term, I hesitate to use it anymore.  But
it would take a much longer message for me to describe what I mean by
Sean Donelan, Data Research Associates, Inc, St. Louis, MO
  Affiliation given for identification not representation

More information about the NANOG mailing list