[Nml-wg] [Nsi-wg] STPs in NSI v2.0

[Roman Łapacz]

W dniu 2012-07-11 17:20, Freek Dijkstra pisze:
> On 11-07-2012 16:39, Roman Łapacz wrote:
>
>>> What do you suggest?
>> I would not use nml namespace if the elements don't apply NML schema.
>> Simply, I would remove it from your examples.
> OK, I got the same feedback in the NSI conference call today.
> I will remove it and leave it up to the NSI-WG what namespace to use,
> but will make it clear that "Port", "PortGroup" and "label" are NML
> constructs.
>
> I'm going to update my proposal to the NSI-WG. What should I use:
>
> Unless you have a strong preference, I'm going to use:
>
>> <sink>
>>      <PortGroup>urn:ogf:network:nordu.net:2012:surfnet-nordunet</PortGroup>
>>      <labelgroup labeltype="http://schemas.ogf.org/nml/2013/10/ethernet/vlan">1800-1899</labelgroup>
>> </sink>
> Instead of:
>
>> <sink>
>>      <PortGroup idRef="urn:ogf:network:nordu.net:2012:surfnet-nordunet"/>
>>      <labelgroup labeltype="http://schemas.ogf.org/nml/2013/10/ethernet/vlan">1800-1899</labelgroup>
>> </sink>

If it's not NML then the first proposal above seems to be better as I 
haven't seen the idRef attribute concept in NSI examples. Use of IdRef 
(as it is in the second example) may raise a question what's for. In 
case of NML/NMC/NM the response is obvious, for NSI probably not.

But to be honest if NSI files include topology information it would be 
ideally if NML proper constructs (nml namespace) were used for that. 
That is why xml namespaces exist to reuse existing solutions.

Roman


> (The reason is that I didn't bother to introduce NML constructs in the
> NSI. In short, I really want to to REFER to NML, but I don't want them
> to BE NML. However, if you really think I should introduce idRef to NSI,
> this is a good time.).
>
>>> I can understand the desire to make it a short piece of NML, but that
>>> would require introducing id/idRef, the hasInboundPort/hasOutboundPort
>>> relations, hence the nml:Relation construct. The original proposal used
>>> query parts in the URN, which is not part of NML (I'm not even sure if
>>> the query part is part of the URN or not -- see my mail
>>> http://www.ogf.org/pipermail/nml-wg/2012-July/001012.html).
>> I like Aaron proposal to use child elements and keep URN
>> identifiers/strings static.
> For those not on the NSI list, Aaron's proposal is here:
> http://www.ogf.org/pipermail/nsi-wg/2012-July/001984.html
> And my proposal to the NSI-WG is here:
> http://www.ogf.org/pipermail/nsi-wg/2012-July/001993.html
>
> I tried to base my proposal on Aaron's examples, with the child element
> for labels. (Although I did change some of the names, and made the label
> syntax a bit more generic to support non-VLAN labels).
>
> Consider this the (over)due credit to Aaron for helping get this work going.
>
> Freek