[core] core management pointers

Juergen Schoenwaelder <j.schoenwaelder@jacobs-university.de> Fri, 07 March 2014 12:35 UTC

Date: Fri, 07 Mar 2014 13:34:51 +0100
From: Juergen Schoenwaelder <j.schoenwaelder@jacobs-university.de>
To: core@ietf.org
Message-ID: <20140307123451.GA69641@elstar.local>
Mail-Followup-To: core@ietf.org
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Disposition: inline
User-Agent: Mutt/1.5.21 (2010-09-15)
Archived-At: http://mailarchive.ietf.org/arch/msg/core/zaCGLi1x76oBJUY_i98y5Z1oBEA
Subject: [core] core management pointers
Precedence: list
Reply-To: Juergen Schoenwaelder <j.schoenwaelder@jacobs-university.de>

Hi,

concerning the management of constrained devices and CoAPs potential
roles in this, let me provide a few pointers. First the two COMAN
documents being discussed in the OPSAWG:

  http://tools.ietf.org/html/draft-ietf-opsawg-coman-use-cases-01
  http://tools.ietf.org/html/draft-ietf-opsawg-coman-probstate-reqs-01

The later one has a new section 1.7 which I think if very useful to
distinguish different configuration and monitoring levels. CoAP based
solutions may want to look at which levels they target.

The NETCONF working group has been rechartered to work on RESTCONF, a
REST-like protocol that provides a programmatic interface over HTTP
for accessing data defined in YANG, using the datastores defined in
NETCONF.

  http://tools.ietf.org/html/draft-bierman-netconf-restconf-04

RESTCONF is not cast into stone yet. It would be very useful to find
out whether RESTCONF can be mapped to CoAP or whether any changes of
RESTCONF would make such a mapping easier.

I also like to repeat what I said at the meeting: what really matters
a lot is reuse of data models. Integration into existing management
systems becomes much easier as long as the naming and the semantics of
management data received via different management protocols is the
same. For monitoring, there are many definitions that can be reused
and MIB data models usually work fine for defining collections of
counters. For configuration, YANG data models are the recommended way
to go these days. But keep in mind that configuration can mean very
different things (section 1.7 in opsawg-coman-probstate-reqs).

I think having a side meeting in Toronto to further discuss CoAP usage
for management and how it relates to RESTCONF and existing data models
and data modeling approaches is a very good idea.

/js

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: +49 421 200 3587         Campus Ring 1, 28759 Bremen, Germany
Fax:   +49 421 200 3103         <http://www.jacobs-university.de/>

[core] core management pointers Juergen Schoenwaelder