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

Re: Proposed Update to Netconf Charter



On Mon, Jul 11, 2005 at 07:51:44AM -0700, David T. Perkins wrote:

> In SNMP, the identification of an instance of management information
> can only be extended in a backwards compatable way through use of
> contexts. In other management systems, such as CMIP and netconf,
> identification of an instance is simply extended by adding a new
> container object. In SNMP operations, a "context" applies to
> all items in the request, response, and notification. However,
> in other management protocols, the added container is needed
> only for those items for which it applies.

Adding a container might be as difficult as adding an index to
a table once a data model has been deployed.

<soap>
Note that CMIP really tried to tackle parts of this problem by
introducing allomorphism. There are a few articles in CMIP Run!
<www.simpleweb.org/standard/iso/cmip-run/> which deal with this
topic in case someone is interested. I am sure Randy knows the
details well and whether this was a success story. In modern 
programming languages such as Java, the concept of interfaces 
became popular and this allows on object to actually support 
multiple interfaces, which can be exploited for backwards
compatibility if you want. But to be honest, I do not see any 
of these fancy concepts applying to XML encoded documents. 
</soap>

I am not advocating to add something like contexts - if you have 
screwed something up in your XML hierarchy, fix it by fixing the 
data model and use name spaces and version numbers to deal with 
backwards compatibility issues. The hope is that the costs of doing 
this and adopting to such changes will be reasonably cheap in XML 
land.  If netconf is successful, we will know in say 5 years whether 
this is actually true. ;-)

/js

-- 
Juergen Schoenwaelder		    International University Bremen
<http://www.eecs.iu-bremen.de/>	    P.O. Box 750 561, 28725 Bremen, Germany

--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>