RE: Policy Statement on Address Space Allocations
I've been reading this address allocation ping-pong match for some time, and I'm a bit confused. I'd like to describe a real-world case, and gain some understanding by getting some advice. On one hand, we have registries that won't hand out /18 class C blocks for fear of depleting the remaining address space too quickly. On the other, we have "core" providers refusing to route longer prefixes because their routers are about to melt down. I'm currently trying to get a new ISP for a small country on its feet. It can't get addresses from its upstream provider, because they've not got enough and are trying to get some more themselves. They also want to be multi-homed from day one. They currently offer all sorts of fully redundant data services, and understandably they can't see why they should have a single point of failure in their global Internet access. We can spell BGP, and we'll teach them how to spell it too. How about some concrete advice on The Right Thing to do from you folks? Thanks, Mathew -- | Mathew Lodge: lodge@houston.omnes.net |5599 San Felipe, 4th Floor | | Internet Specialist, Omnes -- A |Houston, Texas 77056, USA | | Schlumberger/Cable & Wireless company |Phone: +1 713 513 3237 | | http://www.omnes.net/ |Fax: +1 713 513 3126 |
If we believe that we're going to get 1024 multihomable address blocks per /8, then there will not be enough of these for all the providers who can multihome in this NAPatudinous market. IPv4 is going to _force_ us to a multitiered routing system. Everybody, please stop trying to figure out how to make a singly tiered system work out of fear that you will be trapped in a relationship with a bad upstream. Instead, start trying to figure out how to build neutral, lasting upstreams. PCH and CIX are examples, though neither is quite right for various reasons.
participants (2)
-
lodge@houston.omnes.net
-
Paul A Vixie