Pekka Savola wrote:
> 
> On Fri, 21 Dec 2001, Brian E Carpenter wrote:
> > Pekka Savola wrote:
> > > This is getting too implementation-specific, so I guess we had better kill
> > > this thread, at least from here...
> >
> > In fact, it would be a fundamental error to make protocol choices on the
> > basis of perceived implementation glitches in today's popular operating
> > systems. If we do the "right thing", one of these days the o/s will
> > catch up.
> 
> My gripe with 6to4 is that it doesn't discuss the security issues
> properly.  Everything about security is basically written (rather
> concisely), in security considerations, like sugar coating on a cake.

Actually I think the decapsulator pseudocode indicates where checks
should be made, and as far as I know that is the only place where
security would come in. Yes, the description is in boundary condition
terms, which the WG thought was necessary and sufficient at the time.
> 
> IMO, this is the wrong approach.  Security precautions should be discussed
> and handled all the way through the specification (as with Shipworm), and
> in security considerations, a summary and remainder threats discussed.
> Remainder threats are not covered there.

Are you aware of any except spoofing?
> 
> A few points from along the course of this thread:
> 
> - 6to4 does not require the checks and the threats are partially
> downplayed

It's very hard to attach MUST to checks which are implementation
and deployment specific. 

> - automatic tunneling does not discuss the checks even that much
> - issues with multiple use of configured/autotunnel/6to4/... on the same
> box is not covered AFAIK anywhere

Quite true.

> - should autotunnel be deprecated in a more official fashion?

Probably. That means removing it from the address architecture and from
RFC 2893.

> 
> > That doesn't deny the value of informational documents about today's
> > implementation issues, especially security threats.
> 
> This is what my draft was/is partially meant to be; from my perspective (I
> looked at Linux and KAME) it appeared that the checks had been implemented
> only partially or not at all. Perhaps partly because they may not have
> been understood to be important, or were difficult to implement because of
> the problem with multiple mechanisms.

I think your draft is useful.

   Brian


--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to