Re: [Ibnemo] [Sdn] Defining a Common Model for intent

["Natale, Bob" <[e-mail]>]

Apologies to those on the lists who have already seen the attached (Dave Lenrow's introductory presentation to the 2015 Intent Based Network Summit), but I thought that it might be useful for those who haven't.

Avanti,
BobN

-----Original Message-----
From: Ibnemo [mailto:[e-mail]] On Behalf Of STUART VENTERS
Sent: Monday, June 08, 2015 10:22 AM
To: [e-mail]
Cc: [e-mail]; Susan Hares
Subject: Re: [Ibnemo] [Sdn] Defining a Common Model for intent

All,

Over the weekend I figured out what is bugging me about the word intent.

Thinking about the interface (role?) between the customer and service provider.
To provision a service, the provider needs to know the following:
   WHO the customer is. (the billing and technical contact info, perhaps a PO number, and if the customer is a residence or business)
   WHAT sort of service.  (the speeds and feeds, what packet headers the network is aware of, addressing, special things like nat or firewall,  etc)
   WHERE the service should appear. (physical location of the endpoints, maybe logical circuit ID's)
   WHEN the service should work. (Install date, maybe the 100Gig for backups only works at night?)

Conspicuously missing is the following  question:
   WHY the customer wants the service. (Perhaps the provider could provide a better service if he knew this?)

To me, the word 'intent' focuses one to think about WHY.
Saying that the service is provisioned based on intent, implys that the service provider can figure out what to do from mostly the WHY question.
This seems misleading.

That said, to figure out how to specify a service, there are all sorts of interesting questions in the who, what, where, when area.
I wonder if this is where this intent stuff should head?


-Stuart

_______________________________________________
Ibnemo mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/ibnemo