Hi folks,

I am finally trying scratch some time to address IESG feedback related 
to NAT traversal draft. Eric Rescorla (among others) questioned why we 
have chosen to have the locators (aka candidates) in plaintext whereas 
in ICE, the locators are XORred to protect against middlebox tampering. 
The original reasoning for this is was that because that is the way 
non-NAT traversal version of the HIP works (RFC7401).

I don't think we need XORring with HIP because we have more powerful 
mechanisms in HIP. So, I am going to add some text that mandates that 
the LOCATOR parameter must be encapsulated inside ENCRYPTED parameter 
when ICE-HIP-UDP will be used. The tradeoff here is that we favor 
end-host privacy at the cost middlebox transparency.

Please let me know during the two next weeks if you disagree, otherwise 
I consider the issue to be resolved at least from the WG perspective.
_______________________________________________
Hipsec mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/hipsec

Reply via email to