Re: I-D Action: draft-ietf-hip-native-nat-traversal-12.txt

On 6/30/16, 9:06 AM, "Miika Komu" <[email protected]> wrote:
>> Seems like a good idea. No ESP_TRANSFORM -> no need to establish two-way comms between peers.
>> For example, when performing a registration procedure with a relay server.
>
> The direct path could be, of course, used for exchange HIP messages 
> directly (including hiccups v2). Does this make sense?

yes, makes sense

> If not, what should happen when both ESP_TRANSFORM and ICE-HIP-UDP are 
> both negotiated? Or should we just be proactive and state that upon 
> receiving R1, the Initiator MUST NOT include ICE-HIP-UDP if it is not 
> going to employ any ESP_TRANSFORM.

This proposed sentence seems like a good revision.

> Connectivity tests implement the return routability checks. Currently, 
> the NAT mobility triggering mechanism mimics the tree-way procedure in here:
>
> https://tools.ietf.org/html/draft-ietf-hip-rfc5206-bis-12#section-3.2.1
>
> I thought that would nice for implementers but strictly speaking steps 
> 5-6 could skipped since the connectivity checks actually implement the 
> return routability checks.
>
> I can change this if you agree?

Yes, I agree they can be skipped. 
The connectivity checks are more UPDATE packets, and fewer update round-trips seems like faster mobility handover.

-Jeff