[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RRG] anycast/default as a transition tool [On the Transitionability of LISP]



On Fri, Aug 03, 2007 at 08:10:53PM +0300, Christian Vogt wrote:
> Pekka Savola wrote:
> > Maybe we don't need a solution at the edges, just more default routing
> 
> This actually seems to be a very attractive idea.
> 
> Packets destined to IDs, for which no entry exists in the global routing
> table, would by definition be forwarded along a default route until they
> get to a provider that performs the ID-to-locator mapping for the
> destination edge network.

	This is one of the ideas that both Darrel Lewis and Vince
	Fuller have suggested in the past. 

> One disadvantage, though, would be that loops may become difficult to
> prevent unless ALL providers (as of a certain tier) that advertise a
> default route perform the mapping for ALL edge networks using ID space.
> 
> One would also have to think about changes in the Internet business
> model.  As explained in my previous email, a transition towards mapped
> IDs would require an edge network to subscribe to the remote-mapping
> service of as many as possible providers -- in addition to the standard
> subscriptions with the directly connected providers.

	I agree that we need to think (quite a bit) more about
	the/which business models that underly all of this.

	Dave

Attachment: signature.asc
Description: Digital signature