RE: Routing System Scaling - Disaster Looming, but Medium-Term Fixes Known
Bill - If your magical mystical OBGP stuff puts pressure on the global routing system, you can expect at some point that people stop subsidizing your use of their resources, and start filtering or damping you, or charging you real money. If your magical mystical OBGP stuff does not, then it is irrelevant to the scaling of the global routing system by virtue of having no impact on it. There is NO reason why significant site-to-site multihoming need be made visible to the global routing system, if sites are not providing global transit for one another via entities that cannot aggregate the sites behind a single prefix. That UfT and YorkU have some sort of fast direct interconnection just for exchanging traffic between themselves does not need to be known in New York or Amsterdam, or even by ONET. There is probably some opportunity for you to exploit the fact that few networks impose any kind of barrier on resource consumption at the present time. Surely, though, you would be better off looking at the reality that operators face before arguing that your "empower the edge!" ideas are truly cost-saving. Sean.
participants (1)
-
smd@clock.org