In einer eMail vom 20.04.2008 20:45:24 Westeuropäische Normalzeit schreibt [EMAIL PROTECTED]:
One example where, I believe, consideration of economics is due is the question of who should select the ingress link of an edge network (a related thread on this list). Letting the sender of a packet select the remote ingress link (by letting it select the destination transit address) would certainly be an approach that is technically viable, but it would be opposite to the economic view that the receiver, who pays for its ingress link, should be able to control which one is used. - Christian Wrt this example: at first, a routing technology should be provided, that gives you all the options - of course, in a scalable way. For comparison, consider an intradomain-OSPF network topology, and hereby some (destination) node with several links to several neighboring nodes. It can be assumed that at some point in time the rtgwg might provide good solutions to do IP forwarding in compliance with preferencing any particular ingress link at the destination node. Good solutions means, this preferenced choice shall be well balanced wrt to the total traffic to that destination node. However, with the current inter-domain protocols you can forget the pursuing of such goals. Heiner
