Re: [Anima-bootstrap] joiner router stateful/stateless considerations

[peter van der Stok <[e-mail]>]

>  I **SUGGEST** that the anima
> bootstrapping services
>     > both types of networks. That may mean that 2 solutions from the
> ones proposed
>     > in joinrouter document are chosen.
> 
> Which two solutions would you pick?

Below are the options for the managed constrained network, right? Thus 
one of the 2 solutions.

> 
> From the new node to joining router, we have three options:
>      A) CoAP/DTLS, B) HTTPS, C) HTTPS over HTTP proxy.
> 
> I prefer (A) as MTI, with (C) as MAY.

I tend to prefer (A) as I don't understand all consequences of (C)

> From the joining router to the registrar, I prefer method 6, CoAP/DTLS 
> with
> IPIP tunnel.  I think that this works the best in a constrained 
> network,
> where I do not expect to see an ACP.

Sounds reasonable, will try to think of alternatives.

> 
> At the edge of the contrained network, the 6LBR might have an ACP, and 
> would
> further route the join the traffic that way.  The IPIP header might be
> addressed to the 6LBR, and re-encapsulated by the 6LBR as well, but 
> that
> would involve some state (methods 1 or 2) on the 6LBR: That may be 
> acceptable.
> 
> I'm trying also to build a system that could also be initiated from the
> registrar to new node direction, as I've proposed in 6tisch.

I have no preferences on initiaion location for the moment.

> 

> I think that the point is that we don't know (in this document, at this
> stage) precisely if the new node and proxy communicate at layer 2 (e.g. 
> 802.1x)
> or layer 3 (PANA, CoAP/HTTP, etc.).  That's why it says "or"

OK, I got it, a specification "xor" and not an operation "or"

peter