Geographic v. topological address allocation
tli at juniper.net
Fri Nov 21 08:53:33 UTC 1997
> This is true, but the definition of the top of the hierarchy is arbitrary
Not at all true. The top of the hierarchy must be default free.
> and is the nexus of the debate about "topological" versus geographical
> addressing, which I interpret as "ISP at top" versus "exchange point at
> top" hierarchies. Both are valid topological hierarchies.
True, however, geographic addressing has some rather severe practical
problems. The exchange point at the top becomes a single point of
failure. So it needs replication. But then, there needs to be
interconnect between the exchange points. Who provides it?
All this and more has been beaten to death. If you start with the premise
of geographic addressing and try to beat it into working, you end up with
an ISPAC. See ftp://ftp.juniper.net/pub/users/tli/ispac.txt.
More information about the NANOG