> >> I've got some other problems with the document  also - specifically
> >> the idea that link-local addresses are preferable to global addresses.
> 
> > Please let me confirm: are you referring to Rule 2 of source address
> > selection?
> 
> Perhaps you are talking about Rule 8 of destination address selection:
> 
>    Rule 8: Prefer smaller scope.
>    If Scope(DA) < Scope(DB), then prefer DA. Similarly, if Scope(DA) >
>    Scope(DB), then prefer DB.
> 
> This rule may be sometimes controversial, but I personally think it is
> reasonable, because we can (in theory) expect better reachability for
> a smaller scope of destinations.  This is especially true for a
> link-local destination where we don't need L3 routing.

Rule #2 is not a problem, rule #8 is.  The problem is that the address
selected for one connection pair may end up getting reused in other
scope.  So if rule 8 results in an LL or SL address being used, there
might be an attempt to reuse that address (perhaps by another application
component) in a different scope.

There really should be a rule to the effect "avoid scoped addresses
if at all possible".

Keith
--------------------------------------------------------------------
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