Re: [DNSOP] fragmentation itself (Re: FYI: draft-andrews-dnsop-defeat-frag-attack)

[Tim Wicinski <[e-mail]>]

The chairs can work with kazunori on a Monday night discussion (okay, I'll
sign up).

On Mon, Jul 15, 2019 at 10:49 AM Paul Vixie <[e-mail]> wrote:

> On Monday, 15 July 2019 13:35:33 UTC Tim Wicinski wrote:
> > The chairs felt that fujiwara-san's draft was one that needed in person
> > discussion in Montreal.
>
> hopefully kazunori will suggest a bar bof for monday night. i leave
> tuesday
> morning and so will miss the WG meeting. my own concern is more for packet
> size than for fragmentation itself. i'd like to explain my concerns to
> $somebody.
>
> > Also, if folks did not see his presentation at OARC, here are the slides:
> >
> >
> https://indico.dns-oarc.net/event/31/contributions/692/attachments/660/1115/
> > fujiwara-5.pdf
>
> possibly to be argued, but to be not dismissed. as i wrote up-thread, the
> state mass of fragmentation, harshest on receivers but also present on
> transmitters, means the original EDNS0 spec ought to have said "SHOULD be
> of a
> size that allows the full packet with all headers to fit the local
> interface
> MTU and next-hop MTU" and possibly also "MAY set the DF bit if capable of
> hearing and interpreting ICMP messages".
>
> --
> Paul
>
>
>