[Nsi-wg] NSI Topology extensions document

[Henrik Thostrup Jensen]

Hi

On Fri, 20 Dec 2013, Jeroen van der Ham wrote:

>> The schemas/test2.xml is (very) outdated. I suggest removing it.
>
> Done.

Did you forgot to push this?

>> I would - very much - like to have xs:any under the nsi:NSA (I was actually under the impression that it was there already). It is also a bit weird of us to complain about NML not having them, and not adding them ourselves. Here is a patch:
>>
>> --- a/schemas/nsi-ext.xsd
>> +++ b/schemas/nsi-ext.xsd
>> @@ -47,6 +47,9 @@
>>           <xs:element  ref="nsi:Interface" minOccurs="0"  maxOccurs="unbounded"/>
>>           <xs:element  ref="nsi:Relation"  minOccurs="0"  maxOccurs="unbounded"/>
>>           <xs:element  ref="nml:Topology"  minOccurs="0"  maxOccurs="unbounded"/>
>> +          <!-- the local one is needed for xmllint to not complain about determinism -->
>> +          <xs:any      namespace="##local"  processContents="lax"  minOccurs="0"  maxOccurs="unbounded"/>
>> +          <xs:any      namespace="##other"  processContents="lax"  minOccurs="0"  maxOccurs="unbounded"/>
>>         </xs:sequence>
>>       </xs:extension>
>>     </xs:complexContent>
>
> Doesn’t the inheritance from NetworkObject already include the xs:any ?

One should think so. It is certainly possible that it is just 
libxml2/xmllint which is buggy here, but it didn't validate for me. My 
xsd-foo is too weak to really give a proper answer to this.

>> There is a bug in libxml2 that will complain about the model not being 
>> deterministic when not having the local there. It is actually, but 
>> since a lot of people use libxml2, I've added the local element.
>
> Sounds good to me.
>
>> Since we added the labelType in the SwitchingService, I think it should be in the example.
>
> Good point, do you have an example?

Possibly:

     <nml:Topology id="urn:ogf:network:....">
         ...
         <nml:SwitchingService id="urn:ogf:network:example.net:2012:nodeA:switchingService"
                 	      labelSwapping="true"
       			      labelType="http://schemas.ogf.org/nml/2012/10/ethernet#vlan">
         ...

I hope this is right.

>> Will the namespaces by updated?
>
> Yes, the namespace will be updated to the date of publication of the document (as happened with NML). I have not updated them yet, because I have no idea when this would happen.

Ok.

>> One more thing. I think we discussed - and perhaps decided - to an NSI extension for capacity on the ports. AFAICT that isn't it currently; I think we should consider adding it.
>
> Iirc that depended on adding an attribute to a Port, which I’m not sure how to do in an extension schema.

Well, you cannot really define a new type, as the would make something is 
not longer an NML port. I think the way is to define an attribute in 
nsi-ext.xsd , like this:

+
+  <!-- capacity of a port, attribute for Port and PortGroup -->
+  <xs:attribute  name="capacity" />
+

Doing this:

     <nml:PortGroup id="urn:ogf:network:aruba.net:ps-out" nsi:capacity="1000">

Is okay with xmllint at least. But I don't really know if this is the 
correct way of doing things.


     Best regards, Henrik

  Henrik Thostrup Jensen <htj at nordu.net>
  Software Developer, NORDUnet