>> If a BOF is to take place, I'd very much like to see a presentation from a >> potential *customer* for this technology about why they think they need it. > > Fred Baker did include a "customer" presentation (why the Cisco IT > department requires an IPv6 NAT) in his slides at the BEHAVE WG meeting in > MInneapolis.
With all due respect, I was not convinced that the customer use case presented was derived from actual IPv6 experience. Specifically: I fully believe that there's a bunch of experience with the problem in IPv4, as described in draft-baker-v6ops-b2b-private-routing-00, however I was unconvinced that a solution had actually been attempted in IPv6. Furthermore, Fred's approach in this draft is to recommend the use of ULAs, which seems like a perfectly fine solution (the security concerns don't seem fleshed out nor particularly convincing, and I've never been convinced by arguments that devolve into "someone could make a configuration error"). So it seems to me like this draft does *not* count as a case pro IPv6 NAT. It seems to me that ULAs + SPI firewalling (+ optional application-layer gateways, as needed) can be used to solve this particular problem very effectively. I hope I'm not too rude by asking to see the actual customer pain, as opposed to the customer fear, which I consider different. Solving the multihoming problem, though, is still another matter... _______________________________________________ nat66 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nat66
