[Nsi-wg] Rio Inter-Domain Topo Ring v1.1a
Jerry Sobieski
jerry at nordu.net
Fri Sep 2 10:15:53 CDT 2011
Hi John and Atsuko -
Hmmm...yes, the OWL names are a bit confusing...They rdf outout is not
the simple names I tried to use for the SNE graphical layout. What we
currently have is not clear, so we need to address the issue.
For consistency, simplicity, and expediency sake for Rio, I will go
through and change the topo file to use the NSI URN format for network,
stp, and nsa names as used by the CS spec... But I have some concerns
about the long term usefulness or practicality of doing this. But lets
proceed as an interim step to get through Rio - and revisit as part of
the followup topology discusions.
I will do this and circulate a new topo image shortly.
I will also remove the "http....dtox#" prefix - I don't see a value to
that at all in present circumstance.
Jerry
On 9/1/11 10:28 PM, John MacAuley wrote:
> Jerry, Jeroen,
>
> I am a little confused about the names within the topology files. I
> thought we would be using the standard naming schemes we decided upon
> in Salt Lake City. For example, we have a network named
> "http://www.glif.is/working-groups/tech/dtox#Curacao" when we agreed
> they would be named using the URN format
> "urn:ogf:network:nsnetwork: Curacao". Similarly, STPs should be named
> "urn:ogf:network:stp:Curacao:C1" and I would recommend naming NSAs
> using the format "urn:ogf:network:nsa:Curacao-AutoBAHN" or maybe
> "urn:ogf:network:nsa:Curacao:Curacao-AutoBAHN" for completeness. We
> would therefore get the following for an NSNetwork
>
> <owl:NamedIndividual rdf:about="urn:ogf:network:nsnetwork: Curacao">
> <rdf:type rdf:resource="http://www.glif.is/working-groups/tech/dtox#NSNetwork"/>
> <rdfs:comment xml:lang="en">Position : [3690,102]</rdfs:comment>
> <hasSTP rdf:resource="urn:ogf:network:stp:Curacao:C1"/>
> <hasSTP rdf:resource="urn:ogf:network:stp:Curacao:C2"/>
> <hasSTP rdf:resource="urn:ogf:network:stp:Curacao:C3"/>
> <hasSTP rdf:resource="urn:ogf:network:stp:Curacao:C4"/>
> <managedBy rdf:resource="urn:ogf:network:nsa:Curacao:Curacao-AutoBAHN"/>
> </owl:NamedIndividual>
>
> I think this makes it much clearer as to the values someone should use
> within their requests since it now matches the defined formats for the
> protocol elements.
>
> Thank you,
> John.
>
> On 2011-09-01, at 11:08 AM, Jerry Sobieski wrote:
>
>> Hi John-
>>
>> If you use the SNE editor to graphically run a hasSTP link from Aruba
>> to STP A3, it generates correctly. If you run the hasSTP link from
>> the STP A3 to Aruba, it generates the mistake you noted. So I have
>> corrected the topo file.
>>
>> Jeroen- you may want to look at this in the editor. hasSTP should
>> only be allowed to go from NSnetwork object to STP object. We need
>> something else as the reverse relation...maybe "STPof"...?
>>
>> Jerry
>>
>> On 9/1/11 8:04 AM, John MacAuley wrote:
>>> I have not yet gone through the entire file but I did notice this
>>> STP has an STP.
>>>
>>> <!-- http://www.glif.is/working-groups/tech/dtox#A3 -->
>>>
>>> <owl:NamedIndividualrdf:about="http://www.glif.is/working-groups/tech/dtox#A3">
>>> <rdf:typerdf:resource="http://www.glif.is/working-groups/tech/dtox#STP"/>
>>> <rdfs:commentxml:lang="en">Position : [864,558]</rdfs:comment>
>>> <hasSTPrdf:resource="http://www.glif.is/working-groups/tech/dtox#Aruba"/>
>>> </owl:NamedIndividual>
>>>
>>> On 2011-08-31, at 12:54 PM, Jerry Sobieski wrote:
>>>
>>>> Here is an update on the inter-domain topology OWL file (the Rio Ring).
>>>>
>>>> ---> I deleted the intra-domain Ports.
>>>>
>>>> ---> Corrected a mispelled object name.
>>>>
>>>> ---> Added the csProviderContact URL to the NSA object for Aruba
>>>> (OpenNSA), Bonaire (OpenDRAC), and Curacao (AutoBAHN.) Others-
>>>> please forward me a contact place for your NSA.
>>>>
>>>> ---> Added Admin contact info to the NSA adminContact field.
>>>>
>>>> Note- the SNE graphical display only shows a single link between
>>>> objects for connectTo, and mapsTo relations...The RDF is actually
>>>> correct in showing relations both directions, but the editor
>>>> somehow combines them to a single line.
>>>>
>>>> This topology now contains only NSnetwork objects, STP objects, and
>>>> NSA objects.
>>>>
>>>> Each implementation will need to edit this topology file to map
>>>> their STPs to the corresponding internal ports. The mapping is
>>>> done by inserting the string representing the internal port into
>>>> the "mapsTo" field of the corresponding STP. A reverse relation
>>>> should be provided mapping the internal port to the STP. Each
>>>> NSA/network must do this for every STP in their assigned network.
>>>>
>>>> Each network is free to define an internal topology as is most
>>>> convenient for them. The only requirement on internal topology is
>>>> that the STPs defined in the inter-domain topology must map to
>>>> internal ports with the following characteristics: a) 1000
>>>> megabits/second max capacity, and b) untagged.
>>>>
>>>> If your internal topology is static, you can send me your updated
>>>> OWL file and I will integrate the mapsTo fields into my global
>>>> topology file and then subsequent topology updates will have your
>>>> mappings already included...
>>>>
>>>> Tip: The DToX ontology used for the Rio NSI Topology still
>>>> supports the Node, Port, and several other topology objects.
>>>> These can be used by the respective networks to define internal
>>>> topology that is compatible with the inter-domain topology if they
>>>> so choose.
>>>>
>>>> If you need *ANY* help with this at all, explainations, advice,
>>>> questions, etc... PLEASE!!! Let me know! We are up against the
>>>> wall now...simple is better...If I can simplify things for you I am
>>>> glad to do so.
>>>>
>>>> Best regards!
>>>> Jerry
>>>>
>>>> <Rio-Inter-Domain-Topo-Ring-v1.1a.owl>_______________________________________________
>>>> nsi-wg mailing list
>>>> nsi-wg at ogf.org <mailto:nsi-wg at ogf.org>
>>>> http://www.ogf.org/mailman/listinfo/nsi-wg
>>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/nsi-wg/attachments/20110902/39c432f1/attachment-0001.html
More information about the nsi-wg
mailing list