Thank you very much for the interpretation of “Policy Domain”.

Based on the reply from Paul, the term “Policy Domain” in their draft is about 
a “Family (or a group) of Tenants”.
Is it a proper to use “Policy domain” as a term referring to the domain 
applying to a family or a group of tenants? Say a group of Departments 
(tenants) belonging under one organization?

If not, can you suggest a better term?

Thank you.


From: John Strassner []
Sent: Thursday, February 08, 2018 6:08 PM
To: Linda Dunbar <>
Subject: Re: [I2nsf] what does the term "Policy Domain" commonly refer to? (was 
RE: WG Adoption call for

A "Policy Domain" is an administrative domain in which a set of Policies are 
used to ensure that managed entities in that domain behave in a desired manner. 
Policies can be used for configuration, monitoring, access control, and other 

Note that this is a standard term in the academic literature.


On Thu, Feb 8, 2018 at 2:59 PM, Linda Dunbar 
<<>> wrote:

Since you are the policy expert, what does “Policy Domain” commonly refer to?
Can “Policy domain” be one policy applying to a set of tenants? Or one policy 
applying to a set of geographic regions? Or Policy domain being a set of 

Thank you.

From: John Strassner [<>]
Sent: Tuesday, February 06, 2018 5:47 PM
To: Linda Dunbar <<>>
Subject: Re: [I2nsf] WG Adoption call for

IMHO, the purpose of a WG adopting a draft is to acknowledge that the draft is 
a good starting point for the work that WG wants to accomplish. To be perfectly 
clear, I am NOT objecting on the completeness of the document. Rather, I am 
objecting on the technical correctness of the starting point.

I do NOT feel that the proposed documents represent a good starting point. 
Ignoring things that can be easily fixed (e.g., grammar), there are a host of 
problems, such as:
   - what, exactly, is this draft trying to do? I thought I would see YANG for 
policy rules sent over the Consumer-Facing Interface.
     Instead, I see the name of the interface, whose first element is 
multi-tenancy, that also contains policies? Policies do not care
     about multi-tenancy. They do care about domains. The organization of the 
YANG is incorrect.
   - sec 4: in the ieft-i2nsf-cf-interface module
      - why is multi-tenancy at the top of the tree? Shouldn't a DOMAIN be able 
to have multiple tenants?
      - why does a domain have an authentication-method? First, multiple such 
methods should be able to be used. Second, how would a domain know what an 
authentication method even is?
      - why is tenant a sibling of domain, and not a child?
      - why is domain a leaf within policy-tenant? This should be a reference, 
and why doesn't domain have a reference to policy-tenant?
      - policy roles have nothing to do with multi-tenancy - why are they here?

 I could go on, but even the above means that the rest of the YANG will be 

Therefore, the document is NOT a good starting point, and will NOT accelerate 
the path to getting a good RFC.


On Fri, Jan 26, 2018 at 3:23 PM, Linda Dunbar 
<<>> wrote:

The authors of I2NSF Consumer-Facing Interface YANG Data Model

Have requested working group adoption of this draft.

Please bear in mind that WG Adoption doesn’t mean that the draft current 
content is ready, WG Adoption only means that it is a good basis for a working 
group to work on.

While all feedback is helpful, comments pro or con with explanations are much 
more helpful than just "yes please" or "no thank you".

Thank you.

Linda & Yoav

I2nsf mailing list<>


I2nsf mailing list<>

I2nsf mailing list

Reply via email to