Re: Re: proposal for the introduction of Global Identifiers for lightpaths

[Freek Dijkstra <fdijkstr@xxxxxxxxxxxxxx>]

Ronald van der Pol wrote:

> > I like to hear other opinions.
>
> A similar issue was raised in the OGF NML working group:
> http://www.ogf.org/pipermail/nml-wg/2008-September/000175.html
>
> So maybe urn:ogf:... is the way to go.

The identifier in the NML working group are to identify *classes* (e.g. 
"device", "layer", "interface", etc.)

The identifier discussed here are to identify *instances* (e.g. 
sara.nl:lightpath-sara-aarnet-03, 
canarie.ca:lightpath-canet-esnet-080311, etc.)

I would say that identifiers of instances, such as discussed here should 
be part of the namespace of the originator, thus not in a ogf namespace.
So my preference would be to simply use the Internet2 naming scheme 
(GRI), only using a : instead of - as separator, and allowing any string 
as the local identifier (not just numbers). e.g. 
dcn.internet2.edu:lhc-20080923

Basically, I don't say there is a need to append something like 
"urn:ogf:network:lightpath:" before the above GRI (giving 
urn:ogf:network:lightpath:dcn.internet2.edu:dcn.internet2.edu:lhc-20080923), 
since it will (must) be clear from the context that we are taking about 
a lightpath identifier.

This removes the need for the lengthy namespace delegation process (IANA 
has not delegated urn:ogf yet to the OGF).

Regards,
Freek