On Thu, Dec 3, 2009 at 4:30 PM, Dae Young KIM <[email protected]> wrote:
> What every packet has to carry is the node address(IP address), but not the
> session ID. Session ID is a property of the layer up. In the current
> Internet model, socket could be considered as session ID, which, however,
> need not be visible in the packet header where the main job is routing.

DY,

The consequence of that choice is a constraint on the routing system:
because the socket's ID isn't carried in the packet, the IP addresses
carried in the packet must match what the transport protocol expects
or the packet can't be associated with the correct socket. If an
identity for the socket was explicitly included in the packet, that
association function would operate independent of the IP addresses
present... Just as IP routing operates independent of the MAC
addresses attached to the packet.


On Thu, Dec 3, 2009 at 4:34 PM, Scott Brim <[email protected]> wrote:
> No.  I'm saying that routing doesn't care about identification
> functions.

Scott,

To the extent that the identification functions place constraints on
the operation of the routing system, it cares very much. Or if you
want to look at it from the other side: the routing architecture
places constraints on the permissible identification functions and
those constraints have to be adequately specified.

Regards,
Bill Herrin



-- 
William D. Herrin ................ [email protected]  [email protected]
3005 Crane Dr. ...................... Web: <http://bill.herrin.us/>
Falls Church, VA 22042-3004
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to