I suppose it's the same as for any NAT situation, a node trying to send
packets to another in a MAP domain has to know in advance what port to
send to. Otherwise (for shared addresses), the packet is likely to be
routed to the wrong CE. Should the examples mention that the port (e.g.,
port 9030 in example 2) is known in advance to be one that will reach
the desired endpoint?
Is there anything special about MAP compared with the general NAT
situation that makes it easier or harder for senders to learn what ports
they can send to?
Tom Taylor
_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires
