Comment on The why and what of the CIDR Report by Geoff Huston

I remem ber that EGP and GGP were around at that time, and (modulo myrecollections and a not very dense digital record if what we were doing at ther time) NSFNET went with GGP (I have some recollections of the Merrit GateD project?). However, unlike others, I was sitting in Australia in a RIP connection to the NASA Science Internet, so I was not directly involved in the core routing issues in the final days of ARPANET and the collection of US federal agency backbones, so you may well be correct here.

There were lots of "panic points" when local limits were reached - some in the protyocol, some in the router hardware being used at the time. At one point in the very early 90's aussie users could not reach any MILNET sites as I was informed that the MILNET routers had 20K FIB entries max and when that was reached they started pruning out the foreigners! :-)