Re: [netconf] Shepherd review on draft-ietf-netconf-restconf-client-server-26

[Kent Watsen <[e-mail]>]

Hi Tom,

Thank you for your review of the restconf-client-server draft.
Please see below for my responses to your comments.

Kent



> On Jan 3, 2023, at 11:58 AM, tom petch <[e-mail]> wrote:
> 
> Picking a useful looking set of addressees to use but nothing to do with the actual shepherd comments on the I-D

I don't understand this comment.


> I am looking at the I-Ds and have started with restconf; the comments may or may not apply to other I-D, time will tell
> 
> Some are errors, others are matters of opinion which you may not wish to consider at this stage.

Gotcha.


> YANG string unqualified is used as a key for the servers.  I think that this needs some discussion in the I-D pointing out how big a string can be and all the potential characters therein (emoji?),  I see this as a security issue YMMV.

I see the discussion on "netmod".  Juergen recommends a min length, so maybe "length 32..max;" ?


> You have flip-flopped between increases and decreases as the consequence of a periodic connection as opposed to a permanent and both are right and both are wrong until you qualify what resource you have in mind; I think that this needs expanding  in this and other I-D.  Currently it is written as an absolute which is right and wrong.

See my comment in your review of the netconf-client-server draft.


> YANG doctors often point out the redundant elements in identifiers.  Thus a grouping is clearly a grouping where it is defined and where it is used and so ',,,-grouping' is redundant.  

This is more persuasive than similar comments in your review of other drafts.


> I would say the same about connection as in 
> 
> connection-type/connection-type/persistent-connection/persistent
> might be
> connection/type/persistent 

Actually, it's currently just /connection-type/persistent (the "choice" and "case" statements don't appear in the data-tree).


> Arguable '-stack' as well
> 
> The examples use IANA allocated IP addresses - should be ones reserved 
> for documentation 

Now using 192.0.2.2.



> /TPC/TCP/

Fixed.


> IP addresses use the zone format

Indeed.  Does 6991-bis need to define an inet:host-no-zone? 


> 'handled internally'
> internal to what? box, site nework stack,  ... needs a qualification IMHO

Added: "(i.e., not by a TLS-terminator in front of the RESTCONF server)."


> well known port 443 could do with a reference in the YANG, RFC or IANA website

It's in RFC 8040, right?

Kent


> 
> Tom Petch
> 
> 
> 
> ________________________________________
> From: netconf <[e-mail]> on behalf of Kent Watsen <[e-mail]>
> Sent: 20 October 2022 01:00
> To: maqiufang (A)
> Cc: [e-mail]; [e-mail]
> Subject: Re: [netconf] Shepherd review on draft-ietf-netconf-restconf-client-server-26