Re: [nmrg] Fw: Adoption call for draft-li-nmrg-intent-classification-02

[Olga Havel <[e-mail]>]

Hi Jurgen,

Thank you very much for your comments. In regards to solutions, they are the starting points to identify users and different intent types we need to analyse and support. These are the existing solutions from the industry, our proposal is based on information from China Telecom, Huawei and others. The goal was to start from business, from existing solutions provided in the industry where we believe intent driven approach is required (independent of any current implementations, if they are intent driven or not). If some solutions, intent users or intent types are missed, please let us know and we should add them. Or if you think any of those mentioned do not require intent driven approach, please let us know. Your comments would be really welcome.

In regards to starting from intent technologies, it seems you are proposing bottom up methodology, starting from what is available today. That is alternative approach, but in that case we are classifying the baseline and not the target. Our preferred approach was to start classification from the target and not baseline and then position baseline (PoCs/Demos/Products/Applications). 

Best Regards,
Olga




-----Original Message-----
From: Schönwälder, Jürgen [mailto:[e-mail]] 
Sent: Tuesday 10 December 2019 19:39
To: Olga Havel <[e-mail]>
Cc: Ciavaglia, Laurent (Nokia - FR/Paris-Saclay) <[e-mail]>; nmrg <[e-mail]>; nmrg-chairs <[e-mail]>
Subject: Re: [nmrg] Fw: Adoption call for draft-li-nmrg-intent-classification-02

On Tue, Dec 10, 2019 at 07:24:41PM +0000, Olga Havel wrote:
> Hi Laurent,
> 
> Thanks for your comments. Would something like this be better format to communicate our proposed taxonomy?
> 
> [cid:image001.jpg@01D5AF86.26216960]
> 
> 
> But we can also use UML or knowledge graphs as well, the problem is IETF format.
>

Turn it by 90 degrees and you can easily write this down in ASCII.  Or make the 2nd layer boxes simply section titles. The format is seldom the problem.

My problems are terms like "Solution" or "Carrier Solutions" or "Enterprise" - these are extremly bendable terms. Why do we need SD-WAN but not SD-LAN? It seems all arbitrary. Things would be different for me if you can point to concrete intent technology for each and every item in there. Then we get into the direction of a survey plus classification and things start to get useful. If we only brainstrom classifiction terms together, I see little value in this exercise. If you pull together information from concrete intent implementation and you identify and extract common features or design pattern, then this adds value. But yes, this is more work than brainstorming some terms together.

/js

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
Fax:   [phone]         <https://www.jacobs-university.de/>