Re: [tcpm] ICMPv6 Error Handling at TCP draft-fujisaki-tcpm-icmpv6-reaction-00.txt

Hi Fernando,

Thank you very much for your detailed explanation.

 | By translating ICMPv4 codes into ICMPv6 codes you'd basically be 
 | moving ICMPv6 a number of years back in time. The ICMPv6 spec already 
 | includes text that gives you freedom in how to respond to ICMPv6 
 | error messages (e.g., abort a connection in a non-synchronized state 
 | in response to an ICMP error message, or not abort the connection in 
 | response to an error message when the connection is in any of the 
 | synchronized states, or whatever). This freedom already lets you 
 | implement the soft errors behaviour if you want to address the 
 | problem of long delays between connection establishment attempts in 
 | v6 scenarios, as described in the soft errors draft of this WG.

We know this text is in ICMPv6 spec, however, as you know, few IPv6
implementations respond to ICMP error messages currently. (The 9th
slide of below material include the ICMPv6 reaction of current nodes.)

http://www.nttv6.net/~fujisaki/fallback.pdf

We asked some vendors to implement as you suggested, however, they
said they did not because there were no clear definition for that.

 | If you translate the ICMPv4 codes into ICMPv6 codes, then:
 | 
 | * For those scenarios in which the lack of v6 connectivity is 
 | signaled by a so-called "soft error", you won't have solved anything 
 | (as RFC1122 tells you not to abort the connection attempt).

Yes, but in the case that administrators know that, it is possible to
signal some type of so-called "hard error" ICMP, such as admin
prohibit. This can be used in the case that a site uses ULA in their
network.

 | Implementations have been moving away from the strict type/code -> 
 | {hard, soft} error classification since more than ten years ago. And 
 | I honestly think it would be quite a challenge to get the industry 
 | implement an ICMP error processing policy it has already decided not 
 | to implement for ICMPv4.

It looks that some recent OSes still implement ICMPv4 reaction. 

We think we need clear definition of ICMPv6 reaction to ask developers
to implement ICMPv6 error processing.

--
Tomohiro Fujisaki

_______________________________________________
tcpm mailing list
tcpm@ietf.org
https://www1.ietf.org/mailman/listinfo/tcpm