Re: [Detnet] RtgDir Review: draft-ietf-detnet-yang-16

[Don Fedyk <[e-mail]>]

Hi Julien
Thanks for the review.

I have addressed the issues below where noted: Inline [don]
I have placed the pending changes in the public https://github.com/detnet-wg/draft-ietf-detnet-yang  for my co authors to look at and comment. 

The easy way to see the changes is to copy the xml file form git hub and go to the IETF author tools an see the diff. 
Coauthors, please have a look or say that I can push the changes as version 17. 

I will push if my coauthors agree or next week if there is no feedback. 

Thanks
Don   

Subject: RtgDir Review: draft-ietf-detnet-yang-16

Hello,

I have been selected as the Routing Directorate reviewer for this draft. 
The Routing Directorate seeks to review all routing or routing-related 
drafts as they pass through IETF last call and IESG review, and 
sometimes on special request. The purpose of the review is to provide 
assistance to the Routing ADs. For more information about the Routing 
Directorate, please see 
​http://trac.tools.ietf.org/area/rtg/trac/wiki/RtgDir

Although these comments are primarily for the use of the Routing ADs, it 
would be helpful if you could consider them along with any other IETF 
Last Call comments that you receive, and strive to resolve them through 
discussion or by updating the draft.

Document: draft-ietf-detnet-yang
Reviewer: Julien Meuric
Review Date: 2022-09-13
Intended Status: Standards Track


_Summary_

I have some minor concerns about this document that I think should be 
resolved before publication.


_Comments_

The YANG module itself seems almost ready but the text that introduces 
it needs a few clarification or rewording.


_Minor Issues_

- First of all, IdNits points out 3 normative references on 
informational RFCs 8938, 9016 and 9055. Are you sure the 3 of them are 
mandatory to implement the YANG module?

[Don] I think that 8938, 9016 are normative. I have moved 9055.

- In the abstract, I don't really understand the point of the sentence 
"An operator or network controller programs the configuration of the 
devices" since service provisioning on devices along the path is 
previously mentioned. If the intent was to say: " The configuration of 
the devices can be programmed by an operator or a network controller.", 
that feels like stating the obvious for a device-embeded YANG module.

[don] the point was this is a Configured Datapath model not a singling path model.
I've removed the operator /network controller part. 


- In section 4, the wording of "ingress" and "egress" definitions feel 
odd. Is it meant to say "Ingress refers to entering the DetNet 
application layer and egress to leaving the application layer."?
[don]
Old:
All of the layers have ingress/incoming and egress/outgoing operations but any instance may be configured 
as only unidirectional. This means that each unidirectional flow identifier configuration is programmed 
starting at the ingress and flow status is reported at ingress on each end. In the MPLS cases once 
encapsulated, the IP 6-tuple parameters may not be required to be programmed again. In the IP case, without 
encapsulation, various IP flow id parameters must be configured along the flow path.
New:
All of the layers have ingress/incoming and egress/outgoing operations, but any instance may be
configured as only unidirectional. Ingress refers to any DetNet layer where a DetNet context is
applied. Ingress allows functions such as switching, aggregation and encapsulation. Likewise,
egress refers to any DetNet layer where a Detnet context is removed. Egress allows functions
such as switching, disaggregation and decapsulation. This means that each unidirectional flow
identifier configuration is programmed starting at the ingress and flow status is reported at
ingress on each end. In the MPLS cases once encapsulated, the IP 6-tuple parameters may not be
required to be programmed again. In the IP case, without encapsulation, various IP flow id
parameters must be configured along the flow path.



- The described aggregation cases are scoped either as layer N to layer 
N or as layer N to layer N-1. However, there's a relay node case where 
aggregation is described as layer N (forwarding) to layer N+1 (service). 
Since there's no forwarding to forwarding relay case, I suspect a 
mismatch... [Later note: in the model itself, one can find 
"forwarding-to-forwarding aggregation at the ingress node or relay node 
or transit node", so it looks like an issue in the text part.]
[don] Not sure there is a transit node which is forwarding to forwarding and there is a relay node that is forwarding to forwarding.
I struggle to word the missing case?  

- In section 8, the max-loss leaf is an uint32 but is defined as a 
"ratio". Considering the value in the examples (2), it seems that the 
description text (and units?) should be adjusted.
[Don] I think this should be packet loss rate from RFC 9016 and I have adjusted it.
Packet loss rate is a percentage. I have allowed from 0 to 100. With 0.0000000001 being the smallest non zero value. 


_Nits_

------
Abstract
---
s/operational data for DetNet Flows/operational data of DetNet Flows/  
[already 2 "for"s in the phrase]
[Don] Done

------
Section 4.
---
     OLD
Node types typically are logical per DetNet service and one DetNet 
service can be one node type while another is another node type on same 
device.
     NEW
Node types are logical roles per DetNet service: a device along one 
DetNet service can be of one node type, while another service may use 
the same device with a different node type.
[don] Done.

s/edge node(egress/edge node (egress/
s/These may used/These may be used/
s/the configuration need to/the configuration needs to/
s/IP based path/IP-based path/
s/parameters for aggregated flow/parameters for aggregated flow/

[Don] Done

------
Section 10.
---
     OLD
o this also coudl be considered moer sensitive. The trafic profiles liked to
     NEW
so this also could be considered more sensitive. The traffic profiles 
linked to

------
[Don] Done
Regards,

Julien