> -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of Rémi Després > Sent: Tuesday, February 09, 2010 1:14 PM > To: Dan Wing > Cc: [email protected]; 'Templin, Fred L'; 'Behave WG' > Subject: Re: [BEHAVE] [Softwires] PMTU Discovery and ICMPv6 filtering > > Dan, > > There must have been a confusion: I am AGAINST IPv6 PTB > messages with MTU<1280. > (Your last message, addressed to both Fred and me suggests > you believe I support them.)
Sorry if I mis-characterized. > My personal point is AGAINST networks purposely losing > "normal" PTB error messages, i.e. with MTB>1280. > > Considering that these losses become a MAY, even in IPv6, > just because it is believed that some networks lose them > would be IMHO a mistake, at least without a good > understanding of where and why it is done. Remi, I can tell I'm not fully understanding what you're saying. But perhaps if you read Ben Stasiewicz's analysis of ICMPv6 PTB with MTU=1280, you'll be happier -- it shows much higher sucess than his later tests using MTU<1280. His MTU=1280 test is at: http://listserver.internetnz.net.nz/pipermail/ipv6-techsig/2009-October/000708 .html -d > Regards, > > RD > > > > Le 9 févr. 2010 à 21:26, Dan Wing a écrit : > > > >> -----Original Message----- > >> From: [email protected] > >> [mailto:[email protected]] On Behalf Of Rémi Després > >> ... > >> In IPv6, I hope it will be possible to agree at least that > >> PTB messages SHOULD be forwarded (if not MUST). > >> Reasons to not agree should IMHO be seriously documented. > > > > Ben Stasiewicz's tests [*] on the IPv6 Internet show that 56% of > > IPv6 hosts do not properly respond to ICMPv6 packet-too-big > > with an MTU<1280. > > Not related to the point I made. > > > I do not believe additional MUSTs in RFCs > > from the IETF would have a noticeable effect. > > A SHOULD is the proposal. > AFAIK it is not written anywhere that networks are authorized > to do otherwise. > > > > _______________________________________________ > Behave mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/behave _______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
