[Nsi-wg] Pathfinding and Service (a new topology suggestion)

[Henrik Thostrup Jensen]

Hi

On Fri, 27 Feb 2015, John MacAuley wrote:

> Henrik,
> I was talking with Chin about the following example and his understanding was different than mine.  I think I
> have it wrong but wanted to check.
> 
> For the link "urn:ogf:network:aruba:bonaire-otn" my understanding was from Bonaire I could use EVTS to reach
> Curacao with a cost of 2, and I could use OTN to reach Bonaire (back to me) with a cost of 1.

The example is a bit bad as it announces reachability to the network 
bonaire on a link that goes to bonaire (bonaire should have been dominica 
on the reachability for that link).

But service adaptation is implicit and is it not announced where it 
happens.

What it means is:

* The bonaire-otn link supports two services types: EVTS and OTN

For an OTN service on the link, it is possible reach the bonaire network 
(and STP that supports the OTN service in bonaire)

* For an EVTS service on the link, it is possible to reach the bonaire and 
curacao networks (STPs that supports the EVTS service in those networks).


> Chin indicated that the type of service used on the link "urn:ogf:network:aruba:bonaire-otn" is not described
> as it is know between the two ends of the link (implicit), and what the entries indicate is you can reach
> Curacao with EVTS through Aruba at a cost of 2.  So if the implicit service on
> link "urn:ogf:network:aruba:bonaire-otn" is OTN, then adaptation will occur somewhere to get you to the "EVTS"
> service in Curacao.

Mmm... not sure this is correct. Remember, this is chain, and the services 
type is always specified in the request, so there is not such thing as 
implicit service. It is always explicit. The EVTS service might be 
carried on OTN, but that is completely irrelevant.

> So in Chin's explanation a single service type is on the link "urn:ogf:network:aruba:bonaire-otn" and the
> service adaptations are described.  In mine the services on the link are described but adaptation is not.

Adaptation is newer explicitely described. Because there is no technology 
to describe adaptation for. Service adaption does NOT exist in the model. 
I believe your (John's) interpretion is the correct one.

/Henrik

> Could you please clarify?
>
> Thanks,
> John
> 
> On 2015-02-05, at 9:59 AM, Henrik Thostrup Jensen <htj at nordu.net> wrote:
>
>       <Network id="urn:ogf:network:aruba" version="123">
>          <Name>Aruba</Name>
>          <Link id="urn:ogf:network:aruba:bonaire-otn" demarcation="urn:ogf:network:bonaire:aruba-otn">
>              <Name>Bonaire OTN Link</Name>
>              <Service type="EVTS">
>                  <ReachableNetwork id="urn:ogf:network:bonaire" distance="1" />
>                  <ReachableNetwork id="urn:ogf:network:curacao" distance="2" />
>              </Service>
>              <Service type="OTN">
>                  <ReachableNetwork id="urn:ogf:network:bonaire" distance="1" />
>              </Service>
>          </Link>
>          <Link id="urn:ogf:network:aruba:dominica-eth" demarcation="urn:ogf:network:dominica:aruba-eth">
>              <Name>Dominica Ethernet Link</Name>
>              <ServiceTransit type="EVTS" />
>          </Link>
>          <Service type="SwitchingService" />
>       </Network>
> 
> 
> 
>


     Best regards, Henrik

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