[tcpm] On pushback against draft-ietf-tcpm-yang-tcp

["Scharf, Michael" <[e-mail]>]

In today's meeting there was a negative comment on the lack of reference from 
draft-ietf-netconf-tcp-client-server to draft-ietf-tcpm-yang-tcp.

As I have already said orally, I do not think that this concern is valid. In 
particular, such a reference was *not* suggested when I have asked for WG 
adoption in TCPM. Details on WG adoption can be found in my slide deck for 
IETF 108: 
https://datatracker.ietf.org/meeting/108/materials/slides-108-tcpm-draft-scharf-tcpm-yang-tcp-06-02

I don't understand which part of that old presentation would back today's 
comment on the mic.

BTW, as far as I can tell, it is not very common to use TCP-AO to protect the 
*management plane* of a router, i.e., NETCONF, RESTCONF, or SSH. If TCP-AO is 
not used for the app protocol, a YANG model for an app protocol (such as 
draft-ietf-netconf-tcp-client-server) does not need to reference 
draft-ietf-tcpm-yang-tcp with the current scope. That should actually be 
obvious.

Unless I miss something, TCP-AO is typically used to protect *control plane* 
protocols, e.g., BGP, LDP, and the like. As explained before WG adoption, the 
most relevant user of the TCPM model inside the IETF is currently BGP, i.e., 
draft-ietf-idr-bgp-model.

And the document draft-ietf-idr-bgp-model-10 indeeds imports TCP-AO 
definitions from the TCPM document. Granted, the reference to 
"I-D.scharf-tcpm-yang-tcp" needs to be updated to the most recent version of 
the TCPM WG item, but the authors work on that. So, what is missing here?

Michael
(as author who tries to do useful work across different IETF areas)