This is another good reason to publish the information models.  Within the
IETF, it is my experience that data models for new technologies such as the
I2NSF interface get developed quicker if there is a informational model
first. 

 

We did 3 protocol independent models like this in I2RS.  We spread the
logical model to TEAS and it seems to have helped kick start the data model
effort there as well. 

 

Sue

 

From: I2nsf [mailto:[email protected]] On Behalf Of Kepeng Li
Sent: Monday, October 31, 2016 4:21 AM
To: Natale, Bob; John Strassner; Diego R. Lopez
Cc: [email protected]; [email protected]
Subject: Re: [I2nsf] Info models and draft-zhang-i2nsf-info-model-monitoring

 

+1 on separate information model publication.

 

Kind Regards

Kepeng

 

发件人: I2nsf <[email protected]> on behalf of "Natale, Bob"
<[email protected]>
日期: Thursday, 27 October 2016 at 6:11 AM
至: John Strassner <[email protected]>, "Diego R. Lopez"
<[email protected]>
抄送: "[email protected]" <[email protected]>, "[email protected]"
<[email protected]>
主题: Re: [I2nsf] Info models and draft-zhang-i2nsf-info-model-monitoring

 

I agree with John on both points: (1) I2nsf should not ignore non-YANG
domains and (2) therefore, a separate info model document makes very good
sense.

 

Avanti,

BobN

 

From: I2nsf [mailto:[email protected]] On Behalf Of John Strassner
Sent: Wednesday, October 26, 2016 11:41 AM
To: Diego R. Lopez <[email protected]>; John Strassner
<[email protected]>
Cc: [email protected]; [email protected]
Subject: Re: [I2nsf] Info models and draft-zhang-i2nsf-info-model-monitoring

 

Hi all,

 

The main reason for publishing an information model is to support multiple
data models that are derived from it.

 

I realize that the IETF is worried mainly about YANG data models. That is
perfectly well and good. However, security controllers will need to
interface with compute, storage, and other systems that have poor (if any)
YANG coverage. As soon as I go into a data center or a cloud, YANG is simply
not used for compute and storage solutions.

 

This means that we can either ignore that, and build a silo, or take it into
account, or not.

 

   If we choose to ignore, there is no reason to publish the info model.

   If we choose not to ignore, and want to support other data models, then
we SHOULD publish the info model as well. Even if we only have 1 data model
now.

 

Hence, I personally would prefer two publications - one for the info model,
and one for the YANG data model.

 

 

best,

John

 

On Sat, Oct 22, 2016 at 4:29 AM, Diego R. Lopez
<[email protected]> wrote:

Hi Adrian, 

 

Agree: it is a useful tool but should not be a separate publication. The
only reason for publishing the information model could be to do so in the
same document as the data model, as rationale supporting it, and even giving
the opportunity for alternate data models using other data representations
(TOSCA, for example, very much in fashion in cloudspace).

 

Be goode,

 

On 14 Oct 2016, at 11:54 , Adrian Farrel <[email protected]> wrote:

 

Thanks, all, for the useful comments about this document.

It seems clear that there is support for developing this work and producing
a
data model for monitoring.

Two points:

1. As noted by Sue, there is a BoF/WG planned for IETF-98 on "Security
Events".
I suggest you go to that. I will also make sure the AD is aware of the
potential
overlap/interaction.

2. It seems reasonable to me that producing an information model (such as in
draft-zhang-i2nsf-info-model-monitoring) is a useful step toward producing a
data model. I have no objection to using a structured approach. However, my
question about "publication" could be phrased as follows:
- Suppose we decide we want a data model for monitoring
- Suppose we use draft-zhang-i2nsf-info-model-monitoring to guide
  our work on that data model
- Suppose that we push ahead with the data model quite soon so
  that it starts to catch up with the info model
If all of those things apply, why would we need to publish an RFC that
captures
the information model given that we will be publishing a data model shortly
afterwards?
Presumably, once the data model is published, no one will ever read the
information model.
So the information model would be a valuable document working document in
which
the WG would capture its thoughts and consensus, but would be discarded once
the
work to make the data model was complete.

Or am I wrong?

Thanks,
Adrian





-----Original Message-----
From: Susan Hares [mailto:[email protected]]
Sent: 13 October 2016 14:49
To: [email protected]; [email protected]
Subject: RE: [I2nsf] Thoughts on draft-zhang-i2nsf-info-model-monitoring

Adrian:

Why: Monitoring is a key component to I2NSF for monitoring NSF devices.
Monitoring is not the same as NSF devices sending notifications - which is a
push from the NSF devices.  Monitoring may encompasses specific requests to
the device.   Monitoring is different than the DOTS - "help me" cry from a
device under attack.
While I see the security ADs are proposing Security event, it is important
that the I2NSF create monitoring concepts that work with all of the
functions (e.g. querying capabilities, sending/receiving notification, and
events).

Data model versus Information model:  Since we do not seem to have a clear
idea of what the data model should be, it is important to create the
informational models.

The content of the draft is a good first step.

Sue Hares



-----Original Message-----
From: I2nsf [mailto:[email protected]] On Behalf Of Adrian Farrel
Sent: Tuesday, October 11, 2016 5:22 PM
To: [email protected]
Subject: [I2nsf] Thoughts on draft-zhang-i2nsf-info-model-monitoring

Working Group,

Linda and I would like to hear some more from you about
draft-zhang-i2nsf-info-model-monitoring.

Is it something you think we should be working on?
Should we have a separate YANG module for it or fold it into other modules?
If we produce a YANG module, do we still need to publish the information
model?

And, most important, what do you think of the content of the draft?

Thanks,
Adrian

_______________________________________________
I2nsf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2nsf


_______________________________________________
I2nsf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2nsf

 

--
"Esta vez no fallaremos, Doctor Infierno"

Dr Diego R. Lopez
Telefonica I+D
http://people.tid.es/diego.lopez/

e-mail: [email protected]
Tel:    +34 913 129 041 <tel:%2B34%20913%20129%20041> 
Mobile: +34 682 051 091 <tel:%2B34%20682%20051%20091> 
----------------------------------

 


_______________________________________________
I2nsf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2nsf




-- 

regards,

John

_______________________________________________ I2nsf mailing list
[email protected] https://www.ietf.org/mailman/listinfo/i2nsf 

_______________________________________________
I2nsf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/i2nsf

Reply via email to