salo at msc.edu
Tue Jan 28 03:44:45 UTC 1997
> Date: Mon, 27 Jan 1997 17:22:32 -0600 (CST)
> From: Edward Henigin <ed at texas.net>
> Subject: Re: peering charges?
> From what I know, routers (ciscos at least) tend to be
> packet-limited rather than bandwidth limited..
> Isn't it a good enough first approximation to count packets
> rather than sum packet sizes?
It would seem that a reasonable objective for a good network design
would be to ensure that the most expensive or otherwise constrained
component is the bottleneck in the system. Typically, it makes sense to
spend additional money on routers (~ $100,000) to ensure that your links
are kept full (e.g, ~ $200,000 /month for a DS-3).
This might lead you to conclude that you would like to use a measure
that is applicable to your most expensive resource. The bandwidth
used on your major links sounds like a good choice, (estimated by counting
More information about the NANOG