Re: Comments on draft-jinmei-ipv6-rfc2462bis-00.txt

[Brian Haberman <brian@innovationslab.net>]

Jun-ichiro itojun Hagino wrote:

> > Well, there are many networks that are open to the general public, for
> > example wifi networks at airports. 
> > It is true that a bad guy on-link can do a lot of harm, some of which
> > can be alleviated by SEND. However, most of other attacks require a
> > constant stream of packets, and increase the risk that the attack will
> > be detected and traced. The recommendation to turn off the interface
> > amplifies the powers of this bad guy: they can kick someone off the
> > network with a single packet. In short, just because someone broke in,
> > there is no reason to hand her a sledge hammer.
>
> 	but then, if we change it to MAY, what is the point in running DAD
> 	process?  if you do not disable interface (or the address on the
> 	interface) the owner of the same address will get confused,
> 	peers of the address get confused, you will do bad things to the
> 	original owner of the address.  
I see disabling the interface and disabling the address on the interface
as two separate actions.

So, I agree that the interface MAY be disabled.

Brian


--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------