Re: [Detnet] WG adoption poll draft-geng-detnet-conf-yang

Hi Rodney,

Thanks for the detailed clarification!

Please see my response inline...

> -----Original Message-----
> From: detnet [mailto:detnet-bounces@ietf.org] On Behalf Of Rodney
> Cummings
> Sent: Thursday, September 27, 2018 1:01 AM
> To: Lou Berger <lberger@labn.net>; Janos Farkas
> <Janos.Farkas@ericsson.com>; detnet@ietf.org
> Cc: detnet-chairs@ietf.org
> Subject: Re: [Detnet] WG adoption poll draft-geng-detnet-conf-yang
> 
> Thanks Lou,
> 
> > I do think it is fair to sort out the inter-SDO related contents prior
> > to adoption so that we don't send mixed signals to the IEEE or the
> > market.  Can you, either on-list or off-list to authors, state specifically which
> parts of the document/tree you think should be removed?
> 
> I was not clear regarding my question on WG adoption. My concern is exactly
> what you describe above. I have no problem at all with one SDO doing YANG
> work for another SDO's standards, as long as the two SDOs agree to that
> YANG roadmap. I didn't know whether WG adoption of this document
> reflects that sort of inter-SDO agreement.
> 
> My concern is to avoid overlap. If two SDOs develop YANG for the same
> feature, that is potentially painful to product implementers (which includes
> me). I think we can avoid that sort of overlap with communication from both
> sides on this mailing list. I don't think we need formal liaisons or similar.

I can't agree more. 

> 
> Specifically, in draft-geng-detnet-conf-yang, my concern is Section 4 and
> associated YANG.
> I assume that "flow configuration" is intended to be southbound. If it is
> southbound, these nodes directly overlap with YANG in progress for 802.1CB
> (in P802.1CBcv).

The flow configuration in draft-geng-detnet-conf-yang only covers the DetNet flow (layer 3) related parameters. And I assume that P802.1CBcv intends to configure TSN flow(layer 2) related parameters. Why do think they directly overlap? Are you proposing a single model for both DetNet and TSN flow configuration?

> 
> I suppose my higher level question is:
> 	For YANG, which SDO is addressing which topic?

My personal opinion is that IEEE TSN addresses TSN specific configuration, and IETF DetNet WG addresses DetNet specific configuration. For the common part (both apply to TSN and DetNet), more discussions and collaboration between the two SDOs are needed. 

1 ) For flow configuration, since the encapsulations are different, DetNet and TSN should have their own configuration models. 

2 ) For topology discovery: the current DetNet topology model is an augmentation to ietf-te-topology model, it is defined as southbound, may be used as northbound.  Since the bandwidth, PREOF capabilities, queuing delays, queuing algorithms, processing delays etc. parameters are needed for and transparent to both TSN and DetNet. Seems a common model can work for both TSN and DetNet. 

So, for topology model, do we plan to define a common model for both TSN and DetNet? Or define dedicated topology models for TSN and DetNet respectively? 

3) There will be another model: transport QoS model (is not included in the current draft), which may have some common design and parameters for both TSN and DetNet. 

> 
> For example, is IETF DetNet planning to address northbound interfaces?
> 
> If the answer is Yes, then I have no problem keeping the topology content in
> this draft.
> Maybe we will transition that topology content to a northbound-oriented
> draft, but that can be decided later.
> 
> If the answer is Yes, what are we doing with the northbound work that has
> been published thus far in 802.1Qcc?
> 
> Maybe we will transition that northbound work from 802.1 TSN to IETF
> DetNet over time, in order to avoid future overlap. Maybe we will keep that
> topic in 802.1 TSN. Whichever we decide, it would be helpful to make it clear,
> so that we can avoid drafts that overlap unintentionally.

If DetNet plans to define northbound model, I agree with the principle. 

Best regards,
Mach
> 
> Rodney
> 
> From: Lou Berger <lberger@labn.net>
> Sent: Tuesday, September 25, 2018 8:56 AM
> To: Rodney Cummings <rodney.cummings@ni.com>; Janos Farkas
> <Janos.Farkas@ericsson.com>; detnet@ietf.org
> Cc: detnet-chairs@ietf.org
> Subject: Re: [Detnet] WG adoption poll draft-geng-detnet-conf-yang
> 
> Hi Rodney,
> On 9/20/2018 6:29 PM, Rodney Cummings wrote:
> 
> Hi folks,
> 
> At the moment, I am "no / do not support".
> 
> I am confused by what it means for the WG to adopt this draft.
> 
> Adoption of any draft basically means that the WG is formally working on the
> document/topic.  As an individual draft, the draft authors have control of a
> documents contents.  Once a draft is a WG document, control shifts to the
> WG  and the editors/authors are responsible for ensuring a document
> reflects consensus of the WG.  Authors do have flexibility in how they do this.
> For example, proposing changes in on list prior to publication of a new
> version, or simply putting such changes in a new version for review by the
> WG are both common practices.
> 
> 
> The draft contains YANG for configuration of IEEE 802.1 features in a node
> (bridge, router, etc). This includes 802.1CB ("PREOF"), 802.1Qav (credit-based
> shaper), and so on. There are similar YANG projects in 802.1 for these
> features.
> 
> I would expect that there be similarities in configuring DetNet and TSN
> services, but I do agree, we (DetNet and the IETF) should not be defining
> specifics of technology standardized by other bodies as part of our normal
> business.  Now if we synchronize with that other SDO, and they agree that
> we do some complementary work, this is fine -- but I don't believe this is the
> case here.
> 
> 
> 
> Therefore, does adoption of this draft mean that IETF DetNet intends to do
> YANG work that overlaps with 802.1 YANG work?
> 
> I would hope that the answer is No.
> 
> I do think it is fair to sort out the inter-SDO related contents prior to adoption
> so that we don't send mixed signals to the IEEE or the market.  Can you,
> either on-list or off-list to authors, state specifically which parts of the
> document/tree you think should be removed?
> 
> 
> 
> 
> On a different point, I don't think that topology is part of node configuration
> (southbound). Topology is not flow or service related, but it is between the
> Network Operator (controller) and the User... northbound. If I am correct,
> this draft is starting with an incorrect assumption.
> This sounds like a topic that could get sorted post adoption, do you agree?
> 
> Thanks,
> Lou
> 
> 
> Rodney
> 
> From: detnet mailto:detnet-bounces@ietf.org On Behalf Of Janos Farkas
> Sent: Tuesday, September 18, 2018 2:02 PM
> To: mailto:detnet@ietf.org
> Cc: mailto:detnet-chairs@ietf.org
> Subject: [Detnet] WG adoption poll draft-geng-detnet-conf-yang
> 
> Dear all,
> 
> This is start of a two week poll on making draft-geng-detnet-conf-yang-04 a
> working group document. Please send email to the list indicating
> "yes/support" or "no/do not support".  If indicating no, please state your
> reservations with the document.  If yes, please also feel free to provide
> comments you'd like to see addressed once the document is a WG
> document.
> 
> The poll ends Oct 3.
> 
> Thanks,
> János and Lou
> 
> _______________________________________________
> detnet mailing list
> mailto:detnet@ietf.org
> https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__www.ietf.org_mailman_listinfo_detnet&d=DwMDaQ&c=I_0YwoKy7z5L
> MTVdyO6YCiE2uzI1jjZZuIPelcSjixA&r=WA71sf2o7Dw7CbYhFt24DPjt3lJuupsw
> WYdnboKbZ8k&m=2bVCH_eyGhx0Ye6DEdYw9o1jxRMHrpLcmVAKoLkozvI&s
> =cpbNBNL99UFQw8_gmcvSsl_COAbsahIZa0g0oFBSAGE&e=
> 
> 
> _______________________________________________
> detnet mailing list
> detnet@ietf.org
> https://www.ietf.org/mailman/listinfo/detnet