Re: [v6ops] [EXTERNAL] Re: Scope of Unique Local IPv6 Unicast Addresses (Fwd: New Version Notification for draft-gont-6man-ipv6-ula-scope-00.txt)

[Philip Homburg <[e-mail]>]

>The true greatness of dual-prefix multihoming, in my opinion, is
>"hosts get to decide which ISP to use", so a given host can decide
>"I want to use ISP A for my web traffic, and ISP B for my voice traffic"
>(by picking the corresponding source address).

We see that hosts don't even implement the basics, so this seems
to be a bit of a pipe seeam,

>Of course this is not there yet, because there is no guidance *how* to
>pick a proper source address in that scenario (and that leads to the
>other problem of no guidance how to deal with "I picked GUA source A,
>and can't reach my destination, so should I try GUA source B next?").

If you know how to express the diffence in the two upstreams, put it in
Ole's generic RA option, hosts can take it out and use it in
source address selection and you are done. Just a simple matter of
programming.

A generalized Happy eyeballs works fine for unreachable destinations.

>=2E.. and if we can't have the nice things, just using GUA + NPT66 on
>the router will achieve the basic "I want to have two ISPs with failover
>for my web traffic" just fine.

For faulover, you don't need NAT.