Re: deprecating dangerous bit patterns and non-TC non-AXFR

[Mohsen Souissi <[e-mail]>]

 On 20 Aug, Paul Vixie wrote:

 | i worry about
 | <http://tools.ietf.org/html/draft-heffner-frag-harmful-02> and
 | <http://tools.ietf.org/html/rfc4963>.  i mention the
 | earlier-than-final draft because its title was "Fragmentation
 | Considered Very Harmful" which was a play on WRL-87-3
 | "Fragmentation Considered Harmful" (Kent and Mogul, DECWRL 1987).
 | BIND has had to implement all kinds of fallbacks in its EDNS0
 | support due to firewalls that don't allow fragments or which do the
 | wrong thing with them.  i am therefore not as convinced today as i
 | was in RFC 2671 that fragmentation is a reasonable way to expand
 | the DNS message size beyond 512 octets.

==> Hmm, I can see. Otoh, nobody could imagine at that time (and even
10 years ago) how the network-layer would become complex and its
components (often) misbehaving (magic combination of NATs, Firewalls, Load
Balancers...). So don't blame yourself ;-)


[...]

 | there is work going on in this WG to make EDNS0 mandatory.  that will have
 | an impact far broader than what we say to eachother in the forum itself.
 | as you yourself point out:
 | 
 | > 2) It is directly related to the discussion which took place on dnsext ml
 | > while editing dns-protocol-profile:
 | > http://www.ops.ietf.org/lists/namedroppers/namedroppers.2008/msg00037.html
 | > (see section 4.2.2).
 | > 
 | > The I-D expired on Jul 22nd, a copy is available here:
 | > http://ietfreport.isoc.org/idref/draft-ietf-dnsext-dns-protocol-profile/
 | > 
 | > This I-D should be recycled in the near future as I understood from last
 | > dnsext session @ IETF. And then, the result of the current debate may be
 | > helpful for people in charge of editing next version
 | > dns-protocol-profile...
 | > 
 | > Mohsen.
 | 
 | it's possible that you (mohsen) should volunteer to the WG chairs to help
 | edit that document, since i believe its current authors are overcommitted.

==> During IETF@Dublin, Ólafur discussed that matter with the editors
(and maybe others). Stéphane Bortzmeyer (a colleague of mine) was
among them and committed with some others to taking again that task
before next IETF. Hope it will efficiently move forward by then.

Mohsen.


--
to unsubscribe send a message to [e-mail] with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>