Re: [lmap] draft-ietf-lmap-restconf status and way forward

[Benoit Claise <[e-mail]>]

On 6/9/2017 10:14 PM, Juergen Schoenwaelder wrote:
> On Fri, Jun 09, 2017 at 07:50:42PM +0000, Carey, Timothy (Nokia - US) wrote:
>> Juergen,
>>
>> I finally got around to reading the restconf draft (rev-04)
>>
>> In section 4 (RESTCONF as LMAP Report Protocol) - I think you still have some spurious text (would Call Home be relevant between the MA and the Collector) and incorrect use of Controller (should have been Collector)
>> Figure 3 depicts how the connection and the
>> secure transport is established and how reports are delivered to the
>> Controller. Note that it is generally assumed that the Controller is
>> directly reachable from the Measurement Agent. (In situations where
>> this may not be true, RESTCONF Call Home can be used as well but this
>> is not shown here.)
>>
>> If you were trying to say that when the MA issues reports for any reason to the Controller then the above would be correct but I would clearly separate the two scenarios.
>>
> Oops, this is a typing error, this will be fixed in -05.
>   
>> Other than that I am fine with how you describe the Call Home.
> Good.
>   
>> As to the others -
>>
>> Yes informational would be better than standards as there isn't any normative language;
> I will change -05 to say Informational unless the chairs tell me
> otherwise.
Changing to Informational makes sense IMO.

And I agree with Jürgen' statement.

    If the document becomes Informational, I think it is
    better to move the examples inline instead of having them in an
    appendix.


Regards, Benoit
>
>> In section 5 Unless there is specific configuration needs for the MA, Controller or Collector (maybe defining the ports or authentication requirements (e.g., client authentication) as it relates to the RESTCONF client/server interaction that you might want mandate (if so then it should go to standards track BTW).
>> In the TR-069 we specifically called these out so that any Device would know the ports to implement and how a TR-069 agent is identified for authentication as well as what is mandated.
> The ports are defined in the RESTCONF and RESTCONF call home
> specifications. The data model of the authentication configuration
> details is still worked on in NETCONF. It could be that the NETCONF
> work does not detail all aspects of the client side interface but its
> work in progress so we will only know details in a couple of months
> (the new charter says August 2017 but this is an IETF milestone).
>
> /js
>