Mika Liljeberg wrote:
> 
> On Sun, 2003-08-10 at 12:17, Brian E Carpenter wrote:
> > > I would prefer it if the use of semi-unique local scope addresses were
> > > restricted to non-connected networks. For any connected network you can
> > > assume that the network manager is able go to some registry website and
> > > grab a guaranteed unique prefix.
> >
> > Ideally, yes. But that doesn't solve all the real world problems - see
> > the Hain/Templin draft.
> 
> Which requirement are you referring to?
> 
> I don't see any requirement for (even slightly) ambiguous addresses,
> aside from the case where it is, for whatever reason, not possible to
> quickly acquire a guaranteed unique limited range address from a central
> registry.

Exactly! In that case, whatever the reason, a pseudo-random value with a
tiny probability of collision is a fine alternative. 

   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