Hmmm while I understand the reasoning, I do not think that forcing core routers 
to generate ICMP PTB messages is a good idea as this is done by the route 
processor and not in silicon. This would be a nice DoS ;-)

Their ICMP generation is usually rate limited (actually there is a HW rate 
limiter on too-big packets).

OTOH, core should never have a 'small MTU' link, so, this should be only a 
problem in theory.

-éric


> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On Behalf Of
> Havard Eidnes
> Sent: vendredi 6 janvier 2012 09:28
> To: [email protected]
> Cc: [email protected]; [email protected]
> Subject: Re: Fragmentation-related security issues
> 
> >> The problem with RFC4821 (assumming the ICMP-free variant) is
> >> that it has a longer convergnece time that ICMP-enabled PMTU.
> >
> > RFC4821 works even if there are no ICMPs, but will
> > converge more quickly if there are ICMPs. That is why
> > RFC4821 should be a SHOULD for hosts, and generation
> > of ICMPs should be a MUST for routers.
> 
> Does not this also imply that ICMP-generating routers MUST use a
> globally unique IPv6 address as the source of the ICMP?
> 
> Regards,
> 
> - Håvard
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> [email protected]
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------

Reply via email to