Canonical configuration database order

[Andy Bierman <ietf@andybierman.com>]

Hi,

This may be data-model specific, but maybe not.
The SMI uses OIDs to name data, which implies a
canonical order of all possible data (lexi-next in this case).

A canonical order is useful for implementing a diff-config RPC.
CLIs maintain an internal canonical order, even if the user
enters instanced data in a different order (e.g., interface command).

In XML you need to distinguish between multiple named instances
(e.g., a key is defined) vs. multiple unnamed instances (e.g., maxOccurs 
> 1).
The canonical order for named sibling instances is the natural sort order,
based on the key associated with the nodes.  For multiple unnamed
instances, the ordinal position (0 to N-1) uniquely identifies each
node within a set of sibling nodes.  The canonical order of unnamed
instances should not be changed from the PDU order.  However, agent
merge order of unnamed instances is data-model specific.

Is this something of interest to the WG, or something that
every vendor should (potentially) do differently?

Andy






--
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/>