Pekka Savola wrote:

>>> 3. require manual configuration.
> 
> Actually, in some cases (e.g. some ND messages, to prevent spoofing), 
> this wouldn't be all that big an issue :-)


Complete manual configuration in terms of who is expected to use what
level of security is probably out of question; I can't expect to inform
www.cnn.com about my preferences in any manner. However, it may be
possible to think about a single 'flag' that each node has and which
determines who strong security it can handle. But this seems to get
us into one of two undesirable positions:

First, we could allow e.g. www.cnn.com to accept different
levels, but this defeats the purpose of stronger security
since the weaker method could still be used by someone
to trick cnn into diverting your traffic somewhere else.

Second, we could be strict about the levels and only talk to
nodes that use stronger security. But this would restrict us
to a small set of nodes, RO or no RO. Assuming better than RR
security becomes necessary at some point, it's deployment would
be extremely hard (more on this in draft-aura-mipv6-bu-attacks,
section 6.2).

Do you Pekka agree, or did you have some other form of manual
configuration in mind?

Jari

--------------------------------------------------------------------
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