On 2009-03-27 10:44, james woodyatt wrote:
> everyone--
>
> My question about the architectural model of Dave Thaler's SAF being
> related to Realm-specific IP was not an idle one. It's a rare enough
> occasion that I'm the one who remembers something relevant that someone
> like Dave doesn't already know about, that I'm genuinely surprised when
> it happens. Hence, my hair pulling. Sorry about that. (I was also
> over-caffeinated. Sorry about that, too. I was awake until 0300 last
> night fixing bugs.)
>
> Realm-specific IP is described in RFC 3102 and RFC 3103. The protocol
> was even defined to support both IPv6 and IPv4 address realms on either
> [or both] sides of an RSIP gateway.
>
> Framework: <http://tools.ietf.org/html/rfc3102>
> Protocol Specification: <http://tools.ietf.org/html/rfc3103>
>
> It strikes me that if SAF isn't equivalent to RSIP, then I'm not seeing
> the relevant distinctions. Also, if SAF *is* functionally equivalent to
> RSIP, then what makes SAF now more architecturally appropriate than RSIP
> was then? If SAF isn't equivalent to RSIP, then what's the difference
> I'm missing?
Not much, I think, although I'd want to re-read the RSIP documents
again to see if they overlooked any issues that we've learned
about since 2001. Probably what's missing is deployed code.
Brian
_______________________________________________
nat66 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nat66
