> What value is provided by the orthogonality between intra- and inter-domain routing? Independent inter-/intra-domain routing scalability. Note, the current FIB is the combined table for intra- and interdomain next hop forwarding. That's ecaxtly what I have in mind too: Enable next hop lookup by either one table-offset (dest. is in a different geo-patch) or by three table-offsets (dest. is in the same geopatch) - in the end, of course, - no matter whether it is an intra-domain destination or an external destination. Dae, Even if you want to apply Dijkstra, you could do this based on BGP (in this case BGP has just to advertise links/tunnels/LSPs) rather than on a link-state protocol like OSPF. My TARA-topology would be the combination of differently filtered topologies that where disseminated by BGP. E.g. of five topologies (of different zooming levels). The reason is again the too large size of one flat topology. Imagine the number of ASes times the average number of bordering with neighboring ASes. That is too much information which btw isn't needed. As I pointed out before: Given I want to travel from Munich, Karlsplatz, to San Francisco, Golden Gate Bridge, and while I haven't yet entered the plane(link) to San Francisco,international airport, I don't need to see the city map of S.F. nor of any other city I would fly over. That's more than 99,9 % of topology information that I don't need to know. And yet I can find a stretch-1 route to my destination. Heiner -----Ursprüngliche Mitteilung----- Von: Toni Stoev <[email protected]> An: IRTF RRG <[email protected]> Verschickt: Sa., 8. Mai. 2010, 11:09 Thema: [rrg] Inter-domain routing Dae Young, Nice to hear from you. > o Packets would be appended by an AS number of the destination at > exiting its own AS. Packets would need to be globally destined at sending. Destination "AS number + intra-domain locator" is just fine. So destination AS number should be set initially in the packet. This is quite end-to-end, right? > o DFZ routers sees only AS numbers in determining the next hop router. OK. > o The destination IP address would be examined only after the packet > has entered the destination AS. Let's say "locator" instead of "IP address". > RFC1955 by Bob Hinden back in 1996... > Curious to know why this was rejected in IPNG? I'm not curious about past rejections, but if you know there's any good essence for our creativity, say it please. Heiner, > What value is provided by the orthogonality between intra- and inter-domain routing? Independent inter-/intra-domain routing scalability. Are you a proponent? Best regards to you both. Cheers listeners. _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
_______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
