RE: New Maillist for the discussion on the Management of Constrained Networks and Devices

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi All,

as noted in the maillist announcement of IETF secretary "coma" maillist
is for the discussion on the management of constrained networks and
devices. The mailing list will discuss and identify the issues and
requirements and objectives for the management of devices in such an
environment with a special focus on and differentiation of device
classes. 

The idea and trigger for the maillist creation came from a discussion in
the OPS directorate during IETF #82. As
draft-ietf-opsawg-management-stds-07 states IETF so far has not
developed specific technologies for the management of constrained
networks. OPS directorate members stated in IETF #82 that there is a
need to understand the requirements and the necessary solutions for the
management of such a constrained network and its devices. The assumption
people had was that we need a comprehensive management approach to be
able to address the diverse needs of different device classes.

Although the OPS area was doing already standardization work for network
management, the Core WG is one of the essential WGs at IETF interested
in the management of constrained devices. 

Following are some of the questions which have been raised in the OPS
directorate meeting, which are for sure subject to extend from Core WG
pov.:

*	Do we need a new development for IoT management (i.e.
constrained devices) at all? 
-	If yes, what is really needed as standard and what is an
overkill?
*	What type of devices can we support?
*	How are the classes 0-2 for constrained devices defined in
detail?
*	Is some simple configuration management already sufficient?
-	Or do we need also a simple fault management and monitoring?
*	What type of data model modules do we need to standardize? 
-	Just a few core models like ip-cfg, interface?
-	or also other specific models for monitoring?
*	Can we use available management standards and data models as a
starting point and simplify them?
*	Concerning the encoding (JSON, XML, or binary) we seem to be
flexible with tools.
-	Concerning a normative data modeling language, we need to choose
a suitable language capable to prepare structured models. 
-	Is JSON sufficient for this purpose, or should YANG or any other
modeling language be used? 
*	What is appropriate as message transport?
-	CoAP over UDP with soft-transactions?
-	Netconf-Light over TCP?

Obviously the list of the questions above is not exhaustive.

Carsten kindfully provided already in the Prague meeting the definition
of device classes 0-2
(http://www.ietf.org/proceedings/80/slides/core-0.pdf). I think it would
be useful to start a discussion first on the detailed definition of
these device classes 0-2 in constrained networks and based on their
capabilities which functionality they will be able to support. This can
be then used as a guideline for further discussion on the requirements
or objectives for management of such devices. 

As noted in the announcement the result of the coma discussion can lead
to a taxonomy document and a problem statement highlighting the need for
new work.

Please send your opinions/comments to the coma maillist (coma@xxxxxxxx).
To subscribe pls go to: https://www.ietf.org/mailman/listinfo/coma 

Cheers, 
Mehmet 


BTW: Coma has been chosen as the maillist name following the definition
below:

Coma \Co"ma\, n. [L., hair, fr. Gr. ko`mh.]
   1. (Astron.) The envelope of a comet; a nebulous covering,
      which surrounds the nucleus or body of a comet.
      [1913 Webster]




[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]