> I thought the hard part about ensuring convexity wasn't about the routing
> protocol itself, but ensuring convexity in the forwarding of a packet
> with
> dst = global address assigned to site
> src = site local address
I must be missing something. There is really nothing hard about it, as
far as I can see.
When packet comes in from a interface A, you pick scope ids from that
inteface (or just remember where it came from) and get
dst, id-A[global] (well, for global not techinally needed)
src, id-A[site]
Then you forward the packet normally with "dst", and find it lands on
interface B. The only additional test to do is:
if (id-A[site] != id-B[site])
drop packet/or icmp out of scope.
This is clearly explained in scoping architecture, I
thought.
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
