Re: [i2rs] Comments re: draft-i2rs-ephemeral-state-12

["Alexander Clemm (alex)" <[e-mail]>]

Hi Susan,

this looks good - thank you!

Thanks
--- Alex

From: Susan Hares [mailto:[e-mail]]
Sent: Tuesday, July 05, 2016 2:11 PM
To: Alexander Clemm (alex) <[e-mail]>
Cc: [e-mail]
Subject: RE: [i2rs] Comments re: draft-i2rs-ephemeral-state-12

Alex:

Thank you for your comments.    The nit will be released in version 14.


From: i2rs [mailto:[e-mail]] On Behalf Of Alexander Clemm (alex)
Sent: Thursday, June 30, 2016 8:13 PM
To: Susan Hares ([e-mail]<mailto:[e-mail]>)
Cc: [e-mail]<mailto:[e-mail]>
Subject: [i2rs] Comments re: draft-i2rs-ephemeral-state-12

Hi Susan,

this document looks very good & I clearly support it.

Just two very minor comments:


-          editorial nits - page  2:  "Sections 7" --> "Section 7"; "is I2RS protocol requirement" --> "is an I2RS protocol requirement"



-          I consider Ephemeral-REQ-03 as very important ("may have constraints that refer to operational state").  I am wondering, should the draft mention how to deal with the fact that it is possible for operational state to dynamically change.  I would think it is might be worth stating something to the effect that constraints should be assessed when ephemeral state is written, and that situations are conceivable where violations of such constraints might occur due to changing of operational state after the write occurred.   By the nature of the issue, the framework must allow for that; how to deal with such a situation and maintain integrity of the ephemeral configuration in such cases is up to the client.



Alexander:


   Ephemeral-REQ-03: Ephemeral state may have constraints that refer to
   operational state, this includes potentially fast changing or short
   lived operational state nodes, such as MPLS LSP-ID or a BGP IN-RIB.



Is this your suggested change?


   Ephemeral-REQ-03: Ephemeral state may have constraints that refer to
   operational state, this includes potentially fast changing or short
   lived operational state nodes, such as MPLS LSP-ID or a BGP IN-RIB.
   Ephemeral state constraints should be assessed when the ephemeral
   State is written, and if the constraint change after that time
   the I2RS agent should notify the I2RS Client.

If this expresses your desired change, please let me know - I will add it to version 14.




One thought re: section 9, clearly we have a requirement to support subscriptions against ephemeral data; is there a requirement for subscriptions to be ephemeral themselves?  (I think it is implicitly supported via dynamic subscriptions.)

Yes, the subscriptions for the ephemeral only models must be ephemeral.   Is this text addition ok?

Pub-Sub-REQ-03:   The subscription service must support subscriptions which are ephemeral.
(E.g. An ephemeral data model which has ephemeral subscriptions.)



Thanks
--- Alex

Thank you for your comments.

Sue