Well, I don't think it's worth getting into a discussion about whether NAT-PT is required to have per-flow state or SIIT is forbidden from having per-flow state, so I'll just observe that, unless there's also port mapping going on, NAT-PT mostly just deals with addresses. There is an address mapping table, but it's not per-flow. The only exceptions to this are for application protocols that need ALGs.
In the case of the NAT-PT implementation I did a few years ago, we didn't need port mapping (v4 edge site, v6 network core), so we didn't keep per-flow state in the NAT-PT core at all, and the ALG stuff was handled in separate per-application-protococol modules that hooked in via connection hijacking (so-called "transparent proxy", not that I like with the name). The ALGs had to fetch the address mapping data out of the NAT-PT core, but that's trivial. Anyway, the point of the above is just that most traffic through this engine (SMTP, HTTP, SSH, TFTP, NTP, ...) was following the SIIT rules, except for using a different mapping algorithm between the IPv4 and IPv6 addresses. -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
