|> Where I have N source addresses and M |> destination addresses, this is easily shown to be |> O(N*M). | |Well, not quite, if you have an address selection algorithm |that excludes some combinations up front. Also, do we expect |N or M to be >3 in many cases, or even >2 in most cases? | |So I think the practical value will be less than you fear, |typically 4, and >9 would be very rare. Not that this is |negligible, but it's not unthinkable either. | |draft-ietf-shim6-locator-pair-selection and |draft-ietf-shim6-failure-detection talk about example mechanisms |for this, by the way.
This is exactly the problem I am interested in. - Shall the routing system be responsible for handling all cases? - Prohibits the routing system hosts to select the exit path to DFZ? - How is the host informed what SA is optimal? Probing or signaling? BRDP provides locator liveness information to hosts with path metrics. Hosts select one or two SA and probe (e.g. first responder on TCP SYN). Probing is O(M) or 2*O(M). Caching on M helps reducing probing overhead. And a name resolving system may (should ?) help in selecting the best M. Mobility or SA-swap after ISP link failure is to be solved by another mechanism, BRDP is not the complete solution. It is the missing piece of the jigsaw puzzle. Other mechanisms work fine with BRDP (MIP, SHIM6, HIP, and all translator and map&encap solutions). I still believe in the end-to-end principle and the cooperative model ! Teco. _______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
