>
>From: Reinaldo Penno <[email protected]>
>To: "Vijay K. Gurbani" <[email protected]>, Ben Niven-Jenkins
> <[email protected]>
>Cc: alto <[email protected]>
>Subject: Re: [alto] ALTO Interoperability test cases (draft)
>Message-ID: <ca2dfdd4.48a91%[email protected]>
>Content-Type: text/plain; charset="us-ascii"
>
>I second having 1:1 relationship. We explored having an IP address in more
>than one PID during the protocol design and there were some difficult
>cases.
>
>Having said that, a _private_ IPv4 can and would certainly appear in more
>than one PID if these represent different VRFs. It goes back to some of
>the
>gaps identified in penno-cdn
>
>Thanks,
>
>reinaldo
Minor pedantic point: I believe "1-1" would mean one IP address per PID. I
think you're really talking about a 1:N relationship -- one PID, N
addresses.
I just assumed a CIDR was in only one PID, and hence the
longest-prefix-match rule ensures that an IP address can only be in one
PID.
Private addresses & virtual private nets open up a whole can of worms. But
I think the only safe way to deal with that is to say that if an ALTO
server deals with private addresses, then the server and the clients must
be inside that private net. Then it becomes a non-problem.
- Bill Roome
_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto
