From: Edwards, Andrew M (andrew.m.edwards@boeing.com)
Date: Tue Aug 23 2005 - 17:13:18 GMT-3
So I was playing with this little feature last night and came into a
little bit of a quandry... mostly because the docCD is so good at
explaining the implementation side of their services...
Anyhow, so here is my observation and questions:
The v4v6 source and v6v4 source address static translations seem to
apply to both source or
destination address. I verified this with show ipv6 nat translations .
I know that this is meant to have a v6 host talk to a v4 host in their
"native" protocol via translations of 6-4 and vice versa.
What I dont understand is how the v6 destination address (route) is
known to the v6 source; and similarly how the v4
destination address (route) is known to the v4 source.
Or, put another way, a v6 sourceA asks for a v6 service via dns. It gets
a response with a v6 destinationB address.
So the flow is: sourceA - destinationB (all v6 native)
Now, the NAT-PT router must somehow advertise the destinationB ipv6
address into the V6 domain... how?
Same problem exists for version 4 return traffic.
Does anyone know how these issues are resolved?
Perhaps someone here knows of a working example. I was able to lab up a
NATPT router doing a static translation, but I had to specify
Both the v6v4 and v4v6 static translations. Does someone have a working
example of a dynamic translation either one way as the doccd says or
both?
Andy
This archive was generated by hypermail 2.1.4 : Sun Sep 04 2005 - 17:01:20 GMT-3