The Great Exchange

Jerry Scharf scharf at
Sat May 30 00:03:00 UTC 1998

> Why would it need to be done at the backbone?
> Isn't it possible to have a statistically valid sampling of packets
> even at the highest speeds?  You'd know better than I would.  I hope
> the answer is yes, because that is valuable data even if not used for
> billing.
> -- 
> Shields, CrossLink.

And when someone argues that your data and/or collection methods are flawed 
and refuses to pay, what do you do? The circuit switch world that people are 
comparing this to keep full records at all network ingress/egress points, 
which between ISPs are often their busiest boxes. AT&T does not trust PacBell 
to just send them the money, both sides account each transaction. What is 
valuable for engineering and what is required for accounting and billing are 
way different, one has to be able to stand up in court.

The other point of unmetered usage is a direct motive to keep the growth of 
use as high as possible. Some people believe that the "meteoric rise in 
Internet usage" is an important part of marketing plans and thus things that 
reduce it are bad. I have certainly used arguments like this in the past to 
argue against metered networking in corporations. You may kill far more than 
wasteful use if you discourage people from using the Internet whenever they 
want to, which metered billing would.


More information about the NANOG mailing list