[BEHAVE] Comments on draft-bagnulo-behave-nat64-00: NAT

[Magnus Westerlund <[e-mail]>]

Authors,
(As individual)

I read your document and I kind of reacted to that you call the 
translation process being like SIIT. To me it seems not to fulfill the 
one to one mapping that SIIT does and not touching ports. So I think you 
are much more in the region of doing things similar to a IPv4-IPv4 NAT 
with the IPv6 being the private side. You will have to deal with all the 
gory details of transport protocol translations and their affect on ICMP.

I am at least one that think when specifying these devices also should 
include support for our newer transport protocols. Especially as there 
is relatively little that blocks them in the IPv6 world compared to 
IPv4. And if initiating from IPv6 then you can expect most servers to 
have public IPv4 address that may actually be reachable over DCCP or 
SCTP. So not specifying how to handle these protocols will be simply one 
more roadblock against there usage and a better future.

I get the impression that your device will at least be performing 
address independent mapping allowing current STUN and ICE based or 
similar NAT traversal to work. However, when it comes to filtering, is 
really the NAT64 the place to perform any filtering. Isn't it much 
simpler and requiring less state to be non-filtering when it comes to 
mapping and let the receiving host deal with any unwanted traffic?

Then there is the mapping lifetime question. You specify the maximum 
lifetime to 5 minutes for UDP. What timer length where you considering 
for TCP?

Have you made any considerations on what traffic directions that will 
keep the timer alive?

Cheers

Magnus Westerlund

IETF Transport Area Director & TSVWG Chair
----------------------------------------------------------------------
Multimedia Technologies, Ericsson Research EAB/TVM
----------------------------------------------------------------------
Ericsson AB                | Phone [phone]
Färögatan 6                | Fax   [phone]
S-164 80 Stockholm, Sweden | mailto: [e-mail]
----------------------------------------------------------------------
_______________________________________________
Behave mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/behave