Gentlefolks,

I note that Gadi Evron was, until recently, employed by Afilias, the
same company as Joe Abley.  At present, acccording to another recent
NANOG controversy, Mr. Evron.  Mr Hankins is also not an independent
source, being part of ISC, Joao Damas' (document author)  employer.

Also, I do not think that we need to require aristotelian proof. The
basis of my objection isn't the lack of aristotelian proof.  Rather, it
seems reasonable to require _some_ evidence that this is a real problem,
especially in light of the contrived and exaggerated nature of the
claims; the fact that there are DNS attacks that are easier to conduct;
the fact that the alternative attack doesn't risk detection; the fact
that the alternative attack is harder to mitigate; as well as the
previously discredited source(s) of the claims [See
http://www.iadl.org/nanog/nanog-story.html and
http://www.iadl.org/maps/maps-story.html]

Given that the questioned sources form only a tiny part of even just the
North American ISPs, it shouldn't be very hard to find credible
sources---that is, if indeed this is a real problem that is widely
experienced by internet service providers and that this problem is
serious enough to justify the costs of closing open recursors.  But so
far, We've seen no direct evidence nor any indirect evidence.  
Anecdotes and personal assurances from a tiny group that has
collaborated (properly and improperly) in the past is insufficient to
justify the costs of implementing this change. 

I am also reminded of another point that hasn't been brought up
recently: BCP38 provides a complete and general solution for this and
other spoofing attacks.  Given BCP38, there is really no need for this
document.  BCP38 should protect many services that could potentially be
abused by spoofing, including the legitmate uses of open recursors.  
The efforts spent on this document (both in writing and in later
implementation) would be better applied to promoting and implementing
BCP38.

I might suggest a poll of ISPs, and if 5000 or so ISPs worldwide agree
that open recursors attacks are a current, serious problem that can't be
solved by BCP38, then its a problem that should be acted on.  However,
given past experiences with blacklists (particularly the proponents
association with disreputable blacklists), we should take care that the
proponents do not unduly solicit or threaten ISPs to obtain agreement.

Thanks,

                --Dean

On Fri, 5 Sep 2008, David W. Hankins wrote:

> [For brevity, this is intended as a message in support of Joe's
>  position.  I think my original got eaten in the earlier mail
>  server event announced on ietf@, so apologies for any duplicates.]
> 
> On Tue, Sep 02, 2008 at 03:46:48PM -0400, Joe Abley wrote:
> > My point is that there are a large number of distributed denial of  
> > service attacks happening every day, on a scale large enough to  
> > involve multiple providers and cross-organisational teams for  
> > mitigation.
> 
> For informational purposes, I'd like to point out that yesterday on
> the NANOG mailing list, it was asserted that DNS Amplification attacks
> are being observed by one security worker (Gadi Evron) on a seemingly
> daily basis, frustrated by the lack of adoption of BCP 38 (which is
> proposed as the root cause). [1]
> 
> 
> Let me say that it is entirely right to suggest that in this case, if
> you are engaged in a dialogue of logical deduction, then in the face
> of the claim that something does not exist, the responsibility of
> argument is to prove that thing does exist, on the basis that one
> cannot reasonably prove non-existence of any physical object (or
> event) with Aristotelian tenacity.
> 
> Which is problematic because such a proof (with Aristotelian tenacity)
> in this case would require publishing of normally witheld and guarded
> data in provably unaltered forms.  This may not even be possible.
> 
> This would appear then to be an impasse if the IETF required such
> tenacity.
> 
> Fortunately, the IETF works on a basis of consensus among
> practicioners, not on a basis of Aristotelian deductive proofs of
> draft contents and volunteers' opinions.  I'm content to agree with
> the other WG participants that DNS Amplification attacks do persist in
> the modern day, and that it is useful to write and publish a document
> that seeks mitigation.
> 
> I hope that the WG's consensus will be so measured by the chairs.
> 
> 
>  [1] - http://www.merit.edu/mail.archives/nanog/msg11131.html
> 
> 

-- 
Av8 Internet   Prepared to pay a premium for better service?
www.av8.net         faster, more reliable, better service
617 344 9000   



_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to