Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Ignas: 

Thank you for the clarifying email.  I just wanted to understand the type of 
discuss.   This statement clarifies the discuss: 

  DISCUSS is based on the impact of proposed mechanism to operations

This matches the following category from DISCUSS:"It would present serious 
operational issues in widespread deployment, by for example neglecting network 
management or configuration entirely." 
 
May I suggest if the TE is a minor issue, you move the TE comparison to 
comments so we can focus on the others?  You can always raise it to the DISCUSS 
level later after we solve the main issue.   

Do the suggestions that Yan start to address the problem? Is she heading in the 
right direction? If not, we may need to open up the discussion to a wider 
audience in the WG and the I2RS AD. 

Cheerily, Susan Hares 



When you say "not of primary importance", does this mean the I2RS/TE models are 
in the DISCUSS or out of it? 



-Original Message-
From: Ignas Bagdonas [mailto:ibagd...@gmail.com] 
Sent: Thursday, April 5, 2018 4:36 PM
To: Susan Hares; war...@kumari.net; martin.vigour...@nokia.com
Cc: i2rs@ietf.org; draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; 
i2rs-cha...@ietf.org; 'The IESG'
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on 
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

Hi Sue,

I am responding as a DISCUSS holder. Warren may have different ABSTAIN views.

On 05/04/2018 16:12, Susan Hares wrote:
> Ignas and Warren:
>
> Your comments on the IESG telechat (4/5/2018) have two components:
> 1) a DISCUSS based on direct comparison of the I2RS models (dynamic 
> and configuration) with the TE models (configuration only),

This is incorrect. DISCUSS is based on very different background. I2RS and TE 
models are a small fragment in that background, and as I have mentioned in 
previous mails, it is not of primary importance.

> 2) "This will not work".

This is incorrect. DISCUSS is based on the impact of proposed mechanism to 
operations.

Please take a look at my DISCUSS notes again. And let's analyse them in detail 
then.

Network management mechanisms do not exist in vacuum, they are applied onto the 
entities being managed. Management mechanisms need to fit with the entities (in 
a broad sense of the word, including network elements and how those elements 
bind together - the network design) that they are trying to manage. Network 
design prescribes how the management mechanism will look and operate. DC fabric 
topology document appears to try to do the opposite - it provides assumptions 
and restrictions on how the network design that would be suitable for this 
model to be applied to has to look like. And that design appears to be 
substantially different than majority of practical deployments in DC network 
space. Given that this document is on standards track, it sends a strong 
message to the community saying that if one wants to use IETF approved DC 
network management mechanism, one needs to design a network based on the 
assumptions in this document. This is the core of my DISCUSS - the document as 
it is
  worded now prescribes network designs and manageability approaches that are 
disconnected from operational reality. 
If this was an informational document - likely I would not have much concerns 
on the same basis, but STD track document has a different impact.

Now onto the detailed parts of the DISCUSS. The scope of this model - where is 
it intended to be applied? Underlay? Overlay? Both at the same time? The text 
intermixes underlay and overlay scope, there are parts that seem to target one, 
the other, and both of them. The scope of ODL FAAS project does not help either 
- it intermixes the concepts even more freely. What might have been authors' 
intention - and I am not prescribing that, just guessing - was to take a single 
overlay instance and present it as an underlay, a "fabric as a service" in 
marketing terms, and to provide a model for managing the elements of that 
specific instance _ONLY_. I do not see problems with such approach, but the 
document needs to be very clear on the scope and the proposed mechanisms. If 
authors' scope intentions were different then it needs to be clarified in 
detail - the document, at least to me, does not provide clarity on the scope. 
This is a major part of DISCUSS.

The clarity of the model's logic. Model based network management is the new 
command line interface. It needs to be no worse than the industry norms for 
CLI. Having in a model leaf-this with a description of "this is leaf-this" does 
not seem to be sufficient. This is a minor part of DISCUSS.

TE topology - I fail to see anything resembling suggestions in the text of the 
DISCUSS. That is a question. It is a minor part of DISCUSS.

Stages, ports, roles, policies - those all are questions. All are minor parts 
of DISCUSS.

RPC - again, that was a question. I fail to see anything resembling a 
suggestion. 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Ignas Bagdonas]

Hi Sue,

I am responding as a DISCUSS holder. Warren may have different ABSTAIN 
views.


On 05/04/2018 16:12, Susan Hares wrote:

Ignas and Warren:

Your comments on the IESG telechat (4/5/2018) have two components:
1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
configuration) with the TE models (configuration only),


This is incorrect. DISCUSS is based on very different background. I2RS 
and TE models are a small fragment in that background, and as I have 
mentioned in previous mails, it is not of primary importance.



2) "This will not work".


This is incorrect. DISCUSS is based on the impact of proposed mechanism 
to operations.


Please take a look at my DISCUSS notes again. And let's analyse them in 
detail then.


Network management mechanisms do not exist in vacuum, they are applied 
onto the entities being managed. Management mechanisms need to fit with 
the entities (in a broad sense of the word, including network elements 
and how those elements bind together - the network design) that they are 
trying to manage. Network design prescribes how the management mechanism 
will look and operate. DC fabric topology document appears to try to do 
the opposite - it provides assumptions and restrictions on how the 
network design that would be suitable for this model to be applied to 
has to look like. And that design appears to be substantially different 
than majority of practical deployments in DC network space. Given that 
this document is on standards track, it sends a strong message to the 
community saying that if one wants to use IETF approved DC network 
management mechanism, one needs to design a network based on the 
assumptions in this document. This is the core of my DISCUSS - the 
document as it is worded now prescribes network designs and 
manageability approaches that are disconnected from operational reality. 
If this was an informational document - likely I would not have much 
concerns on the same basis, but STD track document has a different impact.


Now onto the detailed parts of the DISCUSS. The scope of this model - 
where is it intended to be applied? Underlay? Overlay? Both at the same 
time? The text intermixes underlay and overlay scope, there are parts 
that seem to target one, the other, and both of them. The scope of ODL 
FAAS project does not help either - it intermixes the concepts even more 
freely. What might have been authors' intention - and I am not 
prescribing that, just guessing - was to take a single overlay instance 
and present it as an underlay, a "fabric as a service" in marketing 
terms, and to provide a model for managing the elements of that specific 
instance _ONLY_. I do not see problems with such approach, but the 
document needs to be very clear on the scope and the proposed 
mechanisms. If authors' scope intentions were different then it needs to 
be clarified in detail - the document, at least to me, does not provide 
clarity on the scope. This is a major part of DISCUSS.


The clarity of the model's logic. Model based network management is the 
new command line interface. It needs to be no worse than the industry 
norms for CLI. Having in a model leaf-this with a description of "this 
is leaf-this" does not seem to be sufficient. This is a minor part of 
DISCUSS.


TE topology - I fail to see anything resembling suggestions in the text 
of the DISCUSS. That is a question. It is a minor part of DISCUSS.


Stages, ports, roles, policies - those all are questions. All are minor 
parts of DISCUSS.


RPC - again, that was a question. I fail to see anything resembling a 
suggestion. That is a minor part of DISCUSS.




I need clarity to guide the WG/authors to a successful resolution of your 
DISCUSS/Abstain.

1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
configuration) with the TE models (configuration only).


The DISCUSS problem is above the specific details of specific models.


The I2RS models are both available for the dynamic and configuration datastore. 
 The dynamic configuration models are for models that do not exactly align with 
perceptions of the configuration model.These set of models are for 
situations that do not align with the configuration store only.

As Martin indicated, trying these alternate  management Yang models in 
dynamic/configuration model configuration needs feedback based on deployment 
and interoperability issues.

This does not align with RFC8342 (NMDA) or and I2RS requirement documents 
(RFC8242).If this is your DISCUSS Criteria, then I have a strong objection 
to your DISCUSS based on these RFCs.

If the Discuss/Abstain is based on one of the following discuss criteria, 
please state this clear in your emails so I can guide the authors and the WG.



1) The protocol has technical flaws that will prevent it from working properly, 
or the description is unclear in such a way that the reader cannot understand 
it without ambiguity.


Yes. Model is not a 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Warren:

Thank you for making clear your concerns.   I started with the same concerns 
and came to resolve these concerns. I will try to write up a clear summary of 
why I resolved it.  However, I do not have some of the same datacenter 
perspective you or Ignas has.  I would like your review on these concerns and 
the authors comments on these points.  

The delay until Friday is due to my day jobs (professor and other IETF hats).  
I'm just out of time to do a proper job of this work. 

Sue Hares 

-Original Message-
From: Warren Kumari [mailto:war...@kumari.net] 
Sent: Thursday, April 5, 2018 1:52 PM
To: Susan Hares
Cc: Ignas Bagdonas; martin.vigour...@nokia.com; i2rs@ietf.org; 
draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; i2rs-cha...@ietf.org; 
The IESG
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on 
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

On Thu, Apr 5, 2018 at 11:12 AM, Susan Hares  wrote:
> Ignas and Warren:
>
> Your comments on the IESG telechat (4/5/2018) have two components:
> 1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
> configuration) with the TE models (configuration only), 2) "This will not 
> work".  I need clarity to guide the WG/authors to a successful resolution of 
> your DISCUSS/Abstain.
>
> 1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
> configuration) with the TE models (configuration only).
>
> The I2RS models are both available for the dynamic and configuration 
> datastore.  The dynamic configuration models are for models that do not 
> exactly align with perceptions of the configuration model.These set of 
> models are for situations that do not align with the configuration store only.
>
> As Martin indicated, trying these alternate  management Yang models in 
> dynamic/configuration model configuration needs feedback based on deployment 
> and interoperability issues.
>
> This does not align with RFC8342 (NMDA) or and I2RS requirement documents 
> (RFC8242).If this is your DISCUSS Criteria, then I have a strong 
> objection to your DISCUSS based on these RFCs.
>
> If the Discuss/Abstain is based on one of the following discuss criteria, 
> please state this clear in your emails so I can guide the authors and the WG.
>
> 1) The protocol has technical flaws that will prevent it from working 
> properly, or the description is unclear in such a way that the reader cannot 
> understand it without ambiguity.
> 2) Widespread deployment would be damaging to the Internet or an enterprise 
> network for reasons of congestion control, scalability, or the like.
>
> These are objections I take seriously as a shepherd/WG chair.  However, I 
> need you both to disambiguate between these two components.   I have been 
> trying to get clarity on which DISCUSS criteria Ignas' comments indicate.
>
> I expect an IESG members to be able to inform a WG chair which the discuss 
> criteria you are specifying.

I specifically did not ballot DISCUSS as there are no Discuss criteria which I 
felt applied - Abstain is a non-blocking position (and so practically 
indistinguishable from No Record, and fairly similar to No Objection).

Much of my twitchiness with the document is that it feels like it is trying to 
define what a datacenter fabric is (in an Introduction and a single term - 
"Fabric: also known as a POD, is a module of network, compute, storage, and 
application components that work together to deliver networking services.  It 
represents a repeatable design pattern.  Its components maximize the 
modularity, scalability, and manageability of data centers.").

There are a huge number of different data center "fabric" designs, and I don't 
really think that the document goes into enough detail about what exactly it is 
meaning when it says fabric. It also asserts:
"for a DC network, a fabric can be considered as an atomic structure for 
management purposes." -- at some level this is probably true, but at the level 
that I look at fabrics it definitely isn't -- a fabric is a thingie made up of 
many devices and they all require management and monitoring and care and 
feeding and love and attention.

A lot of the document seems to be written with a specific sets of views and 
assumptions (yes, it is obviously extensible, but the base selections reflect a 
certain understanding / world view):
the 'fabric-type' can be Vxlan, Vlan or Trill (most fabrics I've seen are plain 
ethernet), 'port-types' are "Port type: ethernet or serial or others" (I've 
never seen a serial fabric, and this leaves out InfiniBand), 'traffic-behavior' 
can be normal or policy-driven (with no description of what the actual 
difference is, nor what 'policy-driven' means), the 'fabric-options' are not 
really options I'd expect to be of primary import, the 'gateway-mode' options 
of 'centralized' vs 'distributed' don't really align with the descriptions (on 
spine vs leaf), nor do I 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Warren Kumari]

On Thu, Apr 5, 2018 at 11:12 AM, Susan Hares  wrote:
> Ignas and Warren:
>
> Your comments on the IESG telechat (4/5/2018) have two components:
> 1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
> configuration) with the TE models (configuration only), 2) "This will not 
> work".  I need clarity to guide the WG/authors to a successful resolution of 
> your DISCUSS/Abstain.
>
> 1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
> configuration) with the TE models (configuration only).
>
> The I2RS models are both available for the dynamic and configuration 
> datastore.  The dynamic configuration models are for models that do not 
> exactly align with perceptions of the configuration model.These set of 
> models are for situations that do not align with the configuration store only.
>
> As Martin indicated, trying these alternate  management Yang models in 
> dynamic/configuration model configuration needs feedback based on deployment 
> and interoperability issues.
>
> This does not align with RFC8342 (NMDA) or and I2RS requirement documents 
> (RFC8242).If this is your DISCUSS Criteria, then I have a strong 
> objection to your DISCUSS based on these RFCs.
>
> If the Discuss/Abstain is based on one of the following discuss criteria, 
> please state this clear in your emails so I can guide the authors and the WG.
>
> 1) The protocol has technical flaws that will prevent it from working 
> properly, or the description is unclear in such a way that the reader cannot 
> understand it without ambiguity.
> 2) Widespread deployment would be damaging to the Internet or an enterprise 
> network for reasons of congestion control, scalability, or the like.
>
> These are objections I take seriously as a shepherd/WG chair.  However, I 
> need you both to disambiguate between these two components.   I have been 
> trying to get clarity on which DISCUSS criteria Ignas' comments indicate.
>
> I expect an IESG members to be able to inform a WG chair which the discuss 
> criteria you are specifying.

I specifically did not ballot DISCUSS as there are no Discuss criteria
which I felt applied - Abstain is a non-blocking position (and so
practically indistinguishable from No Record, and fairly similar to No
Objection).

Much of my twitchiness with the document is that it feels like it is
trying to define what a datacenter fabric is (in an Introduction and a
single term - "Fabric: also known as a POD, is a module of network,
compute, storage, and application components that work together to
deliver networking services.  It represents a repeatable design
pattern.  Its components maximize the modularity, scalability, and
manageability of data centers.").

There are a huge number of different data center "fabric" designs, and
I don't really think that the document goes into enough detail about
what exactly it is meaning when it says fabric. It also asserts:
"for a DC network, a fabric can be considered as an atomic structure
for management purposes." -- at some level this is probably true, but
at the level that I look at fabrics it definitely isn't -- a fabric is
a thingie made up of many devices and they all require management and
monitoring and care and feeding and love and attention.

A lot of the document seems to be written with a specific sets of
views and assumptions (yes, it is obviously extensible, but the base
selections reflect a certain understanding / world view):
the 'fabric-type' can be Vxlan, Vlan or Trill (most fabrics I've seen
are plain ethernet),
'port-types' are "Port type: ethernet or serial or others" (I've never
seen a serial fabric, and this leaves out InfiniBand),
'traffic-behavior' can be normal or policy-driven (with no description
of what the actual difference is, nor what 'policy-driven' means),
the 'fabric-options' are not really options I'd expect to be of primary import,
the 'gateway-mode' options of 'centralized' vs 'distributed' don't
really align with the descriptions (on spine vs leaf), nor do I
understand what that would be critical enough that it is given such
priority and things like the base protocol used isn't),
the descriptions for things like 'fabrictypes:node-ref' ("The device
the fabric includes.") are very unclear.


The document / model is at such an unusual level of abstraction
(simultaneously very high and low) that I simply don't understand how
this could usefully be used -- if I view a "fabric" as simply a
commodity thingie that I ride over, and the model is simply
informational, then it contains much which is not useful to me, and
leaves out things like cross sectional BW / locality hints, etc which
would be. If I instead view this as something that I might want to use
to administer / monitor / model a fabric then it is way too
simplistic.


But, again, I think that this is simply that my views / levels of
abstraction don't align with whatever the authors / WGs are -- and so
I'm not blocking the document, 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Ignas and Warren:

Your comments on the IESG telechat (4/5/2018) have two components: 
1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
configuration) with the TE models (configuration only), 2) "This will not 
work".  I need clarity to guide the WG/authors to a successful resolution of 
your DISCUSS/Abstain. 

1) a DISCUSS based on direct comparison of the I2RS models (dynamic and 
configuration) with the TE models (configuration only). 

The I2RS models are both available for the dynamic and configuration datastore. 
 The dynamic configuration models are for models that do not exactly align with 
perceptions of the configuration model.These set of models are for 
situations that do not align with the configuration store only.   

As Martin indicated, trying these alternate  management Yang models in 
dynamic/configuration model configuration needs feedback based on deployment 
and interoperability issues. 

This does not align with RFC8342 (NMDA) or and I2RS requirement documents 
(RFC8242).If this is your DISCUSS Criteria, then I have a strong objection 
to your DISCUSS based on these RFCs. 

If the Discuss/Abstain is based on one of the following discuss criteria, 
please state this clear in your emails so I can guide the authors and the WG. 

1) The protocol has technical flaws that will prevent it from working properly, 
or the description is unclear in such a way that the reader cannot understand 
it without ambiguity.
2) Widespread deployment would be damaging to the Internet or an enterprise 
network for reasons of congestion control, scalability, or the like.

These are objections I take seriously as a shepherd/WG chair.  However, I need 
you both to disambiguate between these two components.   I have been trying to 
get clarity on which DISCUSS criteria Ignas' comments indicate. 

I expect an IESG members to be able to inform a WG chair which the discuss 
criteria you are specifying. Please let me know if you have additional 
questions.  Your comments on the telechat did not indicate you understood my 
concerns. 

Cheerily, Susan Hares


-Original Message-
From: Susan Hares [mailto:sha...@ndzh.com] 
Sent: Wednesday, April 4, 2018 6:08 PM
To: 'Ignas Bagdonas'; 'The IESG'
Cc: i2rs@ietf.org; draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; 
i2rs-cha...@ietf.org
Subject: RE: [i2rs] Ignas Bagdonas' Discuss on 
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

Ignas:

I'm not saying yes/no to change this model.  I've forwarded Yan's comments 
regarding changes to your specific issues.  Yan is very proactive.  It is 
likely that she will change most of the details if you respond to her.

I am acting as a shepherd/WG chair.  I'm trying to determine  Discuss Criteria 
are you using from the following document are you using to restrict I2RS models 
(dynamic + configuration capable) because there are TE specific models in the 
same area:
https://www.ietf.org/blog/discuss-criteria-iesg-review/

This goes against the design criteria I2RS has used.

You asked for text sequences that lead me to this conclusion:
(1st thread):
 [Ignas] > Excellent. Please get feedback from user community - even if it is 
not yet implemented and operations groups will not be able to provide feedback, 
architecture and engineering groups look into upcoming things and will have 
what to say.

[Sue]: (Repeating earlier comments from email and shepherd's  )  We obtain a 
vendor (Huawei) and a target deployment situation (Data Centers) with two 
potential data centers in China who wanted to work with this type of logical 
deployment.   To this shepherd's eyes, this is the operational information.

[Sue] (new clarifying comments): If you are still objecting, what else do you 
want as proof that the WG did due diligence on obtaining the operational 
feedback for deployment of a model which has I2RS capabilities (dynamic and
configuration) must be judged against any TE (configuration only ) data models.

(2) Further indication that you are comparing I2RS data models against the 4 TE 
models without a consideration to dynamic datastore design issues:

[Sue's comment]:  3rd - How many of the user community have implemented I2RS 
dynamic models or the RFC version of the TE models?

[Ignas] I am aware of 1 I2RS model family implementation. I am aware of 4 TE 
model implementations that I happen to use daily, and aware of several more 
that I do not deal with. And I am not certain what value such counting of 
implementations brings to this discussion.

Summary of my question:
I2RS models (configuration and dynamic) are different than regular TE models 
and you are lumping the I2RS models in with the configuration datastore TE 
models.  Please provide me with the DISCUSS criteria or RFCs you are using to 
make categorization.

After we settle on these issues, we can go on to the other comments.

Sue Hares

-Original Message-
From: Ignas Bagdonas 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Ignas:

I'm not saying yes/no to change this model.  I've forwarded Yan's comments 
regarding changes to your specific issues.  Yan is very proactive.  It is 
likely that she will change most of the details if you respond to her. 

I am acting as a shepherd/WG chair.  I'm trying to determine  Discuss Criteria 
are you using from the following document are you using to restrict I2RS models 
(dynamic + configuration capable) because there are TE specific models in the 
same area:  
https://www.ietf.org/blog/discuss-criteria-iesg-review/

This goes against the design criteria I2RS has used. 

You asked for text sequences that lead me to this conclusion: 
(1st thread): 
 [Ignas] > Excellent. Please get feedback from user community - even if it is 
not yet implemented and operations groups will not be able to provide feedback, 
architecture and engineering groups look into upcoming things and will have 
what to say.

[Sue]: (Repeating earlier comments from email and shepherd's  )  We obtain a 
vendor (Huawei) and a target deployment situation (Data Centers) with two 
potential data centers in China who wanted to work with this type of logical 
deployment.   To this shepherd's eyes, this is the operational information.   

[Sue] (new clarifying comments): If you are still objecting, what else do you 
want as proof that the WG did due diligence on obtaining the operational 
feedback for deployment of a model which has I2RS capabilities (dynamic and 
configuration) must be judged against any TE (configuration only ) data models. 

(2) Further indication that you are comparing I2RS data models against the 4 TE 
models without a consideration to dynamic datastore design issues: 

[Sue's comment]:  3rd - How many of the user community have implemented I2RS 
dynamic models or the RFC version of the TE models?

[Ignas] I am aware of 1 I2RS model family implementation. I am aware of 4 TE 
model implementations that I happen to use daily, and aware of several more 
that I do not deal with. And I am not certain what value such counting of 
implementations brings to this discussion.

Summary of my question: 
I2RS models (configuration and dynamic) are different than regular TE models 
and you are lumping the I2RS models in with the configuration datastore TE 
models.  Please provide me with the DISCUSS criteria or RFCs you are using to 
make categorization.   

After we settle on these issues, we can go on to the other comments. 

Sue Hares 

-Original Message-
From: Ignas Bagdonas [mailto:ibagd...@gmail.com] 
Sent: Wednesday, April 4, 2018 2:40 PM
To: Susan Hares; 'The IESG'
Cc: i2rs@ietf.org; draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; 
i2rs-cha...@ietf.org
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on 
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)



On 04/04/2018 16:03, Susan Hares wrote:
> Ignas:
>
> I am not trying to clarify the specifics.  This response (as I mentioned), 
> will come from the authors.  As a shepherd/WG chair, I am asking for 
> information regarding the basis of your DISCUSS models for specific points.
>
> The 2014 document on the IESG discuss criteria is at:
> https://www.ietf.org/blog/discuss-criteria-iesg-review/
>
> What on this list does the following comment refer to:
> "Why DISCUSS? DC fabric is a type of network topology, yes, it has some 
> specifics, but nothing radically different than any purpose built network 
> topology. Developing a separate model for a specific use case at the same 
> time when there is generic and extensible TE model is questionable."
The fact that for managing similar functionality there appears to be a need for 
different models that would as a result require different model lifecycle 
management clearly falls into the category of operational issues.


> Perhaps you are not considering the fact this is an I2RS model.  Let me 
> provide 3 comments regaring that point:
I am considering the fact that this model defines configuration of something 
that is widely deployed in a way that is not compatible with how it is 
deployed. The fact that this may be I2RS model is not of the primary importance.

>   
> 1st - I2RS is focusing on models that are capable for the dynamic state and 
> configuration state.  These models have qualitative differences.  The 
> mechanisms of a model which does both dynamic state and configuration state 
> is qualitative different that the basic TE models. This model extends the TE 
> models toward this approach (see   module ietf-dc-fabric-topology reference 
> import of ietf-network-topology).
>
> 2nd - During the I2RS process we talked to the TE topology authors and the TE 
> WG.  We agreed that this model has differences.  As a WG Co-chair, I spent 
> time reviewing this interaction.
>
> 3rd - How many of the user community have implemented I2RS dynamic models or 
> the RFC version of the TE models?
I am aware of 1 I2RS model family implementation. I am aware of 4 TE model 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Ignas Bagdonas]

On 04/04/2018 16:03, Susan Hares wrote:

Ignas:

I am not trying to clarify the specifics.  This response (as I mentioned), will 
come from the authors.  As a shepherd/WG chair, I am asking for information 
regarding the basis of your DISCUSS models for specific points.

The 2014 document on the IESG discuss criteria is at:
https://www.ietf.org/blog/discuss-criteria-iesg-review/

What on this list does the following comment refer to:
"Why DISCUSS? DC fabric is a type of network topology, yes, it has some specifics, 
but nothing radically different than any purpose built network topology. Developing a 
separate model for a specific use case at the same time when there is generic and 
extensible TE model is questionable."
The fact that for managing similar functionality there appears to be a 
need for different models that would as a result require different model 
lifecycle management clearly falls into the category of operational issues.




Perhaps you are not considering the fact this is an I2RS model.  Let me provide 
3 comments regaring that point:
I am considering the fact that this model defines configuration of 
something that is widely deployed in a way that is not compatible with 
how it is deployed. The fact that this may be I2RS model is not of the 
primary importance.


  
1st - I2RS is focusing on models that are capable for the dynamic state and configuration state.  These models have qualitative differences.  The mechanisms of a model which does both dynamic state and configuration state is qualitative different that the basic TE models. This model extends the TE models toward this approach (see   module ietf-dc-fabric-topology reference import of ietf-network-topology).
   
2nd - During the I2RS process we talked to the TE topology authors and the TE WG.  We agreed that this model has differences.  As a WG Co-chair, I spent time reviewing this interaction.


3rd - How many of the user community have implemented I2RS dynamic models or 
the RFC version of the TE models?
I am aware of 1 I2RS model family implementation. I am aware of 4 TE 
model implementations that I happen to use daily, and aware of several 
more that I do not deal with. And I am not certain what value such 
counting of implementations brings to this discussion.



See the comments from Chris Hopps and others on slow pace of the netconf work.  
If you are going to restrict to two deployed implementations, you will be 
joining the IDR camp of requirements and slowing the work further.  The only 
reason we require 2 implementations for IDR is for the fragile BGP environment 
and that operators request it due to the global consequences.  Network 
Management of these early yang models have a much more restricted case.


May I ask you to point to where I have said anything about two deployed 
implementations?



Ignas


Sue Hares


-Original Message-
From: i2rs [mailto:i2rs-boun...@ietf.org] On Behalf Of Ignas Bagdonas
Sent: Wednesday, April 4, 2018 10:31 AM
To: Susan Hares; 'The IESG'
Cc: i2rs@ietf.org; draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; 
i2rs-cha...@ietf.org
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on 
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

Hi Sue,


On 03/04/2018 14:59, Susan Hares wrote:

Ignas:

Yan will answer for the authors but I would like to share some information related to the I2RS 
working group reviews.  In your response, please specify why each question is a "DISCUSS" 
quality question rather than a "Comment" question.  The authors and I (as the shepherd) 
will work to resolve both DISCUSS and comment issues.

Let me review only 5 of your many points because they are pointing in a 
direction which is different from earlier QA reviews of this document (rtg-dir, 
ops-dir, yang-doctors) in the 2017-2018 timeframe.
1st - Why TE topology model is not sufficient for modelling the representation 
of DC fabric? Why is DC fabric network topology special compared to any generic 
fabric based topology?

Why DISCUSS? DC fabric is a type of network topology, yes, it has some 
specifics, but nothing radically different than any purpose built network 
topology. Developing a separate model for a specific use case at the same time 
when there is generic and extensible TE model is questionable.


This document was reviewed by authors with the TE topology models to make sure 
there was no conflict or duplication.

Your question implies that only one yang model is appropriate for each type of 
fabric.

That is exactly opposite. What is special about DC fabric that it has to have a 
separate model? What is special about fabric type of topology that it has to 
have a separate model? Why is TE model not suitable?


 This theory of one yang mode per fabric does not apply to dynamic 
(ephemeral) datastore versus configuration datastore models.  It is also not 
true of all models even within the configuration datastore.
Since there is a yang catalog and 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Ignas: 

I am not trying to clarify the specifics.  This response (as I mentioned), will 
come from the authors.  As a shepherd/WG chair, I am asking for information 
regarding the basis of your DISCUSS models for specific points.

The 2014 document on the IESG discuss criteria is at: 
https://www.ietf.org/blog/discuss-criteria-iesg-review/

What on this list does the following comment refer to: 
"Why DISCUSS? DC fabric is a type of network topology, yes, it has some 
specifics, but nothing radically different than any purpose built network 
topology. Developing a separate model for a specific use case at the same time 
when there is generic and extensible TE model is questionable." 

Perhaps you are not considering the fact this is an I2RS model.  Let me provide 
3 comments regaring that point: 
 
1st - I2RS is focusing on models that are capable for the dynamic state and 
configuration state.  These models have qualitative differences.  The 
mechanisms of a model which does both dynamic state and configuration state is 
qualitative different that the basic TE models. This model extends the TE 
models toward this approach (see   module ietf-dc-fabric-topology reference 
import of ietf-network-topology).  
  
2nd - During the I2RS process we talked to the TE topology authors and the TE 
WG.  We agreed that this model has differences.  As a WG Co-chair, I spent time 
reviewing this interaction. 

3rd - How many of the user community have implemented I2RS dynamic models or 
the RFC version of the TE models?  

See the comments from Chris Hopps and others on slow pace of the netconf work.  
If you are going to restrict to two deployed implementations, you will be 
joining the IDR camp of requirements and slowing the work further.  The only 
reason we require 2 implementations for IDR is for the fragile BGP environment 
and that operators request it due to the global consequences.  Network 
Management of these early yang models have a much more restricted case. 

Sue Hares 


-Original Message-
From: i2rs [mailto:i2rs-boun...@ietf.org] On Behalf Of Ignas Bagdonas
Sent: Wednesday, April 4, 2018 10:31 AM
To: Susan Hares; 'The IESG'
Cc: i2rs@ietf.org; draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; 
i2rs-cha...@ietf.org
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on 
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

Hi Sue,


On 03/04/2018 14:59, Susan Hares wrote:
> Ignas:
>
> Yan will answer for the authors but I would like to share some information 
> related to the I2RS working group reviews.  In your response, please specify 
> why each question is a "DISCUSS" quality question rather than a "Comment" 
> question.  The authors and I (as the shepherd) will work to resolve both 
> DISCUSS and comment issues.
>
> Let me review only 5 of your many points because they are pointing in a 
> direction which is different from earlier QA reviews of this document 
> (rtg-dir, ops-dir, yang-doctors) in the 2017-2018 timeframe.

> 1st - Why TE topology model is not sufficient for modelling the 
> representation of DC fabric? Why is DC fabric network topology special 
> compared to any generic fabric based topology?

Why DISCUSS? DC fabric is a type of network topology, yes, it has some 
specifics, but nothing radically different than any purpose built network 
topology. Developing a separate model for a specific use case at the same time 
when there is generic and extensible TE model is questionable.

> This document was reviewed by authors with the TE topology models to make 
> sure there was no conflict or duplication.
>
> Your question implies that only one yang model is appropriate for each type 
> of fabric.

That is exactly opposite. What is special about DC fabric that it has to have a 
separate model? What is special about fabric type of topology that it has to 
have a separate model? Why is TE model not suitable?

> This theory of one yang mode per fabric does not apply to dynamic 
> (ephemeral) datastore versus configuration datastore models.  It is also not 
> true of all models even within the configuration datastore.
> Since there is a yang catalog and selection of yang models is specific to a 
> implemented, there has been no early winnowing of the yang models per type.  
> If you are insisting on this theory of "one yang model" per fabric type, 
> please provide an RFC reference so that I can help review this DISCUSS 
> criteria with the authors.
>
> This yang model has been implemented by 1 vendor, and there was interest by 
> other vendors.  A deployment target has been identified for this model, and 
> feedback is expected from the users.
Excellent. Please get feedback from user community - even if it is not yet 
implemented and operations groups will not be able to provide feedback, 
architecture and engineering groups look into upcoming things and will have 
what to say.

Speaking of implementations, the ODL faas project (from where the majority 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Ignas Bagdonas]

Hi Sue,


On 03/04/2018 14:59, Susan Hares wrote:

Ignas:

Yan will answer for the authors but I would like to share some information related to the I2RS 
working group reviews.  In your response, please specify why each question is a "DISCUSS" 
quality question rather than a "Comment" question.  The authors and I (as the shepherd) 
will work to resolve both DISCUSS and comment issues.

Let me review only 5 of your many points because they are pointing in a 
direction which is different from earlier QA reviews of this document (rtg-dir, 
ops-dir, yang-doctors) in the 2017-2018 timeframe.



1st - Why TE topology model is not sufficient for modelling the representation 
of DC fabric? Why is DC fabric network topology special compared to any generic 
fabric based topology?


Why DISCUSS? DC fabric is a type of network topology, yes, it has some 
specifics, but nothing radically different than any purpose built 
network topology. Developing a separate model for a specific use case at 
the same time when there is generic and extensible TE model is 
questionable.



This document was reviewed by authors with the TE topology models to make sure 
there was no conflict or duplication.

Your question implies that only one yang model is appropriate for each type of 
fabric.


That is exactly opposite. What is special about DC fabric that it has to 
have a separate model? What is special about fabric type of topology 
that it has to have a separate model? Why is TE model not suitable?



This theory of one yang mode per fabric does not apply to dynamic 
(ephemeral) datastore versus configuration datastore models.  It is also not 
true of all models even within the configuration datastore.
Since there is a yang catalog and selection of yang models is specific to a implemented, 
there has been no early winnowing of the yang models per type.  If you are insisting on 
this theory of "one yang model" per fabric type, please provide an RFC 
reference so that I can help review this DISCUSS criteria with the authors.

This yang model has been implemented by 1 vendor, and there was interest by 
other vendors.  A deployment target has been identified for this model, and 
feedback is expected from the users.
Excellent. Please get feedback from user community - even if it is not 
yet implemented and operations groups will not be able to provide 
feedback, architecture and engineering groups look into upcoming things 
and will have what to say.


Speaking of implementations, the ODL faas project (from where the 
majority of this model seems to be coming from) deals with an instance 
of overlay that is subsequently treated as an underlay, and that is 
different that the underlay on top of which that instance is being run. 
If the model focus is on the "fabric as a service" type of topologies 
then it explicitly needs to state that, and then justify why physical 
node properties exist together with logical instance properties in that 
case.




If you are asking this model to cover three-four layer datacenters, this 
approach is opposite some of the initial feedback to the group to keep the 
initial model - that is to keep it simple and restricted to 2 layers in order 
to test the concepts.  If you are asking to provide text (in introduction or 
appendix) that indicates the initial focus, this can be added.
The document as it is written now tries to cover every possible fabric. 
If the scope is intended to be narrower - it needs to be stated. 
Starting from bounded scope is certainly a right thing to do but that is 
not how the document reads now.




2nd - Multiple layers and multiple roles.
Why DISCUSS? Two stage fabrics and fabrics with a perfectly clean node 
role separation do indeed exist, but that is not necessary a common 
deployment model. The document assumes that those are the only possible 
options.



  The authors provide slides in several meetings I2RS meeting repository 
regarding this point.
The initial feedback suggested reducing the "why" text within the draft.   Again, the 
initial feedback was to reduce the initial model's text to 2 layers and simple "whys".  
See proceedings from IETF 95 forward on I2RS on fabric data model for discussions.
Would users of this model also be required to lookup proceedings of past 
IETF meetings in order to understand whether it may fit their use cases?




3rd - The authors will comment on the port restrictions.  Early feedback during 
the I2RS meetings from vendors may have taken the authors down this path.  In 
my review, I expect major issues in this area - but I will let the authors 
comment.
Why DISCUSS? The way how the model specifies port speeds is conflicting 
with common deployment practices.



  4 - policy is simple.

Again, the initial feedback was to keep initial policies simple and gain 
feedback from the deployments.


Why DISCUSS? What kind of policy is being discussed here? The assumption 
of one single universal policy fitting all deployments and use 

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Martin: 

You are correct, this is not a concern about the rules for the Yang module
itself.   Hence, my questions to Ignas about why he placed a DISCUSS based
on "introduction material" that was not specified in any document. 

Thank you for your comments,
Susan Hares 

-Original Message-
From: Martin Bjorklund [mailto:m...@tail-f.com] 
Sent: Tuesday, April 3, 2018 5:57 PM
To: sha...@ndzh.com
Cc: ibagd...@gmail.com; i...@ietf.org; i2rs@ietf.org;
draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org;
i2rs-cha...@ietf.org
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and
COMMENT)

"Susan Hares"  wrote:
> Martin:
> 
> Thank you for the comments on the Yang doctors.  The discussion 
> reference was in the introductory material and not in the descriptions 
> in the YANG text.  Do you also want additional comments in the
introductory section?

No.  The comment was just about the YANG module.

You wrote:

> > Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG suggested 
> > taking out the lengthy descriptions regarding logic and history.  If 
> > we are switching the rules for the YANG models, would you please 
> > update the requirements for the YANG models so that shepherds, 
> > rtg-dir, ops-dir, and yang-doctors can have rules for review clearly 
> > spelled out.

My point is that I don't think we are changing the rules for the YANG
modules, which this reply seemed to indicate.



/martin



> 
> Sue Hares
> 
> -Original Message-
> From: Martin Bjorklund [mailto:m...@tail-f.com]
> Sent: Tuesday, April 3, 2018 4:30 PM
> To: sha...@ndzh.com
> Cc: ibagd...@gmail.com; i...@ietf.org; i2rs@ietf.org; 
> draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org;
> i2rs-cha...@ietf.org
> Subject: Re: [i2rs] Ignas Bagdonas' Discuss on
> draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and
> COMMENT)
> 
> Hi,
> 
> Just a quick comment on the YANG doctor's review.
> 
> "Susan Hares"  wrote:
> > Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG suggested 
> > taking out the lengthy descriptions regarding logic and history.
> 
> It is very common that the YANG doctor review ask for *more* details 
> in the descriptions.  In general, we want the module to have as much 
> explanatory text as possible.  So was the case for the YD review for 
> this document as well; the YD wrote "The descriptions in all YANG 
> Modules are very short/terse."  That was for the -02 version, and even 
> the -00 version did not contain lengthy descriptions AFAICT.
> 
> 
> /martin
> 

___
i2rs mailing list
i2rs@ietf.org
https://www.ietf.org/mailman/listinfo/i2rs

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Martin:

I apologize if my response seem to indicate the descriptions in the YANG
modules or the YANG module overall description.  My response was intended to
query if there were now rules on what needed to be in the introductory
material (non-YANG) in the RFC prior to the YANG module.  

AFAIK, the only formal definition for this portion of the YANG RFCs is
regarding the Yang diagram RFC. 
Therefore, I was asking Ignas to provide clarity on his DISCUSS to determine
references for his expectations for the introductory material.  A DISCUSS
should have some reference to a YANG web page, or an RFC or some written
context to help the shepherd.   As a shepherd unless I have a reference it
is hard to handle conflicting reviews (shorten the introduction vs. add the
introduction). 

If you know of such a web page or RFC, I would appreciate a reference. 

Thank you,
Susan Hares

-Original Message-
From: i2rs [mailto:i2rs-boun...@ietf.org] On Behalf Of Martin Bjorklund
Sent: Tuesday, April 3, 2018 5:57 PM
To: sha...@ndzh.com
Cc: ibagd...@gmail.com; i2rs@ietf.org;
draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; i...@ietf.org;
i2rs-cha...@ietf.org
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and
COMMENT)

"Susan Hares"  wrote:
> Martin:
> 
> Thank you for the comments on the Yang doctors.  The discussion 
> reference was in the introductory material and not in the descriptions 
> in the YANG text.  Do you also want additional comments in the
introductory section?

No.  The comment was just about the YANG module.

You wrote:

> > Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG suggested 
> > taking out the lengthy descriptions regarding logic and history.  If 
> > we are switching the rules for the YANG models, would you please 
> > update the requirements for the YANG models so that shepherds, 
> > rtg-dir, ops-dir, and yang-doctors can have rules for review clearly 
> > spelled out.

My point is that I don't think we are changing the rules for the YANG
modules, which this reply seemed to indicate.



/martin



> 
> Sue Hares
> 
> -Original Message-
> From: Martin Bjorklund [mailto:m...@tail-f.com]
> Sent: Tuesday, April 3, 2018 4:30 PM
> To: sha...@ndzh.com
> Cc: ibagd...@gmail.com; i...@ietf.org; i2rs@ietf.org; 
> draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org;
> i2rs-cha...@ietf.org
> Subject: Re: [i2rs] Ignas Bagdonas' Discuss on
> draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and
> COMMENT)
> 
> Hi,
> 
> Just a quick comment on the YANG doctor's review.
> 
> "Susan Hares"  wrote:
> > Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG suggested 
> > taking out the lengthy descriptions regarding logic and history.
> 
> It is very common that the YANG doctor review ask for *more* details 
> in the descriptions.  In general, we want the module to have as much 
> explanatory text as possible.  So was the case for the YD review for 
> this document as well; the YD wrote "The descriptions in all YANG 
> Modules are very short/terse."  That was for the -02 version, and even 
> the -00 version did not contain lengthy descriptions AFAICT.
> 
> 
> /martin
> 

___
i2rs mailing list
i2rs@ietf.org
https://www.ietf.org/mailman/listinfo/i2rs

___
i2rs mailing list
i2rs@ietf.org
https://www.ietf.org/mailman/listinfo/i2rs

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Martin Bjorklund]

"Susan Hares"  wrote:
> Martin:
> 
> Thank you for the comments on the Yang doctors.  The discussion reference
> was in the introductory material and not in the descriptions in the YANG
> text.  Do you also want additional comments in the introductory section?

No.  The comment was just about the YANG module.

You wrote:

> > Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG
> > suggested taking out the lengthy descriptions regarding logic and
> > history.  If we are switching the rules for the YANG models, would
> > you please update the requirements for the YANG models so that
> > shepherds, rtg-dir, ops-dir, and yang-doctors can have rules for
> > review clearly spelled out.

My point is that I don't think we are changing the rules for the YANG
modules, which this reply seemed to indicate.



/martin



> 
> Sue Hares 
> 
> -Original Message-
> From: Martin Bjorklund [mailto:m...@tail-f.com] 
> Sent: Tuesday, April 3, 2018 4:30 PM
> To: sha...@ndzh.com
> Cc: ibagd...@gmail.com; i...@ietf.org; i2rs@ietf.org;
> draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org;
> i2rs-cha...@ietf.org
> Subject: Re: [i2rs] Ignas Bagdonas' Discuss on
> draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and
> COMMENT)
> 
> Hi,
> 
> Just a quick comment on the YANG doctor's review.
> 
> "Susan Hares"  wrote:
> > Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG suggested 
> > taking out the lengthy descriptions regarding logic and history.
> 
> It is very common that the YANG doctor review ask for *more* details in the
> descriptions.  In general, we want the module to have as much explanatory
> text as possible.  So was the case for the YD review for this document as
> well; the YD wrote "The descriptions in all YANG Modules are very
> short/terse."  That was for the -02 version, and even the -00 version did
> not contain lengthy descriptions AFAICT.
> 
> 
> /martin
> 

___
i2rs mailing list
i2rs@ietf.org
https://www.ietf.org/mailman/listinfo/i2rs

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Martin:

Thank you for the comments on the Yang doctors.  The discussion reference
was in the introductory material and not in the descriptions in the YANG
text.  Do you also want additional comments in the introductory section?

Sue Hares 

-Original Message-
From: Martin Bjorklund [mailto:m...@tail-f.com] 
Sent: Tuesday, April 3, 2018 4:30 PM
To: sha...@ndzh.com
Cc: ibagd...@gmail.com; i...@ietf.org; i2rs@ietf.org;
draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org;
i2rs-cha...@ietf.org
Subject: Re: [i2rs] Ignas Bagdonas' Discuss on
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and
COMMENT)

Hi,

Just a quick comment on the YANG doctor's review.

"Susan Hares"  wrote:
> Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG suggested 
> taking out the lengthy descriptions regarding logic and history.

It is very common that the YANG doctor review ask for *more* details in the
descriptions.  In general, we want the module to have as much explanatory
text as possible.  So was the case for the YD review for this document as
well; the YD wrote "The descriptions in all YANG Modules are very
short/terse."  That was for the -02 version, and even the -00 version did
not contain lengthy descriptions AFAICT.


/martin

___
i2rs mailing list
i2rs@ietf.org
https://www.ietf.org/mailman/listinfo/i2rs

Re: [i2rs] Ignas Bagdonas' Discuss on draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

[Susan Hares]

Ignas:

Yan will answer for the authors but I would like to share some information 
related to the I2RS working group reviews.  In your response, please specify 
why each question is a "DISCUSS" quality question rather than a "Comment" 
question.  The authors and I (as the shepherd) will work to resolve both 
DISCUSS and comment issues.  

Let me review only 5 of your many points because they are pointing in a 
direction which is different from earlier QA reviews of this document (rtg-dir, 
ops-dir, yang-doctors) in the 2017-2018 timeframe.  

1st - Why TE topology model is not sufficient for modelling the representation 
of DC fabric? Why is DC fabric network topology special compared to any generic 
fabric based topology?

This document was reviewed by authors with the TE topology models to make sure 
there was no conflict or duplication.

Your question implies that only one yang model is appropriate for each type of 
fabric.   This theory of one yang mode per fabric does not apply to dynamic 
(ephemeral) datastore versus configuration datastore models.  It is also not 
true of all models even within the configuration datastore. Since there is a 
yang catalog and selection of yang models is specific to a implemented, there 
has been no early winnowing of the yang models per type.  If you are insisting 
on this theory of "one yang model" per fabric type, please provide an RFC 
reference so that I can help review this DISCUSS criteria with the authors. 

This yang model has been implemented by 1 vendor, and there was interest by 
other vendors.  A deployment target has been identified for this model, and 
feedback is expected from the users. 

If you are asking this model to cover three-four layer datacenters, this 
approach is opposite some of the initial feedback to the group to keep the 
initial model - that is to keep it simple and restricted to 2 layers in order 
to test the concepts.  If you are asking to provide text (in introduction or 
appendix) that indicates the initial focus, this can be added. 

2nd - Multiple layers and multiple roles.  

 The authors provide slides in several meetings I2RS meeting repository 
regarding this point. 
The initial feedback suggested reducing the "why" text within the draft.   
Again, the initial feedback was to reduce the initial model's text to 2 layers 
and simple "whys".  See proceedings from IETF 95 forward on I2RS on fabric data 
model for discussions. 

3rd - The authors will comment on the port restrictions.  Early feedback during 
the I2RS meetings from vendors may have taken the authors down this path.  In 
my review, I expect major issues in this area - but I will let the authors 
comment. 

 4 - policy is simple. 

Again, the initial feedback was to keep initial policies simple and gain 
feedback from the deployments. 

5 - You indicate that the document requires a "major" rewrite clarifying the 
logic. 

Earlier feedback (rtg-dir, ops-dir, yang-doctors) on YANG suggested taking out 
the lengthy descriptions regarding logic and history.  If we are switching the 
rules for the YANG models, would you please update the requirements for the 
YANG models so that shepherds, rtg-dir, ops-dir, and yang-doctors can have 
rules for review clearly spelled out. 

Summary on Shepherd's comment:  

The authors will respond to others specifics, but in order to guide these 
diligent authors - I need to know what rules you are setting for the 2018 IESG 
approval of YANG models.  If you are placing a DISCUSS on a YANG model based on 
a set criteria, the criteria needs to be published on a web page or in an RFC. 
If I've missed this criteria that the OPS Area has specified, 

Thank you for your review, 
Susan Hares 

-Original Message-
From: Ignas Bagdonas [mailto:ibagd...@gmail.com] 
Sent: Tuesday, April 3, 2018 7:40 AM
To: The IESG
Cc: draft-ietf-i2rs-yang-dc-fabric-network-topol...@ietf.org; Susan Hares; 
i2rs-cha...@ietf.org; sha...@ndzh.com; i2rs@ietf.org
Subject: Ignas Bagdonas' Discuss on 
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: (with DISCUSS and COMMENT)

Ignas Bagdonas has entered the following ballot position for
draft-ietf-i2rs-yang-dc-fabric-network-topology-08: Discuss

When responding, please keep the subject line intact and reply to all email 
addresses included in the To and CC lines. (Feel free to cut this introductory 
paragraph, however.)


Please refer to https://www.ietf.org/iesg/statement/discuss-criteria.html
for more information about IESG DISCUSS and COMMENT positions.


The document, along with other ballot positions, can be found here:
https://datatracker.ietf.org/doc/draft-ietf-i2rs-yang-dc-fabric-network-topology/



--
DISCUSS:
--

I have concerns about the practical usability of this proposed model as it is 
specified now.

The intended decoupling of fabric implementation