The early days of peering

Jerry Scharf scharf at vix.com
Sat Jun 2 13:45:06 UTC 2001


Sean,

Amusingly, those days did not produce the traffic ratio problem. Everyone 
hauled a circuit from their network to the router in Santa Clara and the 
traffic was symmetric. Thus there was no bandwidth*distance difference in the 
passed traffic.

It was when multiple connection points came in that traffic differential 
became an issue. Those requirements were put in when the perceived center of 
business for the "other ISPs" was local dial connection and bandwidth times 
distance was used to justify the multiple connections, nationwide networks... 
Business changed and now the colos are facing the same arguement for being on 
the opposite side of the spectrum.

jerry





More information about the NANOG mailing list