[Multi-MDS] Added Inter MDS Relationship extension section to content module

[PeterKranich]

📑 Description

• Multi-MDS Feature Part 1: Added Inter MDS Relationship extension section to content module (this PR)
• Multi-MDS Feature Part 2: Define Mulit-MDS scenarios and requirement (tbd)

☑ Mandatory Tasks

The following aspects have been respected by the pull request assignee and at least one reviewer:

• Changelog update (necessity checked and entry added or not added respectively)
  • Pull Request Assignee
  • Reviewer

[d-gregorczyk]

I changed some of the texts. I will also modify your branch and add a proper extension schema plus adaption of the GetMdibResponse.

[JavierEspina]

Do not merge till Stefan feels confident with this PR that he has inherited from Peter.

[JavierEspina]

@PaulMartinsen , there are two requirements defined by this PR (R0613 and R0614). Do they look fine in terms of RI "tagging"?

With the many changes you introduced lately (mostly with PR·469) I am often unsure if we are missing to use any important part of the new RI "tagging" functionality...

[JavierEspina]

@d-gregorczyk , @stefandreher , I have noticed that the Relations extension defined here uses the AbstractMetricDescriptor/Relation defined in BICEPS (and slightly amended in the BICEPS Corrigendum that will be published VERY VERY soon). The usage scope in BICEPS is "represents a relationship between a METRIC and one or more other CONTAINMENT TREE ENTRIEs". However, here it is used to represent relationships between MDSs.

1. Is that stretched usage compatible with BICEPS?
2. And if not (clearly), would it make sense to file a ticket on BICEPS (for the Revision project) to expand the usage scope of AbstractMetricDescriptor/Relation?

[d-gregorczyk]

@d-gregorczyk , @stefandreher , I have noticed that the Relations extension defined here uses the AbstractMetricDescriptor/Relation defined in BICEPS (and slightly amended in the BICEPS Corrigendum that will be published VERY VERY soon). The usage scope in BICEPS is "represents a relationship between a METRIC and one or more other CONTAINMENT TREE ENTRIEs". However, here it is used to represent relationships between MDSs.

1. Is that stretched usage compatible with BICEPS?
2. And if not (clearly), would it make sense to file a ticket on BICEPS (for the Revision project) to expand the usage scope of AbstractMetricDescriptor/Relation?

Actually, I had to introduce a dedicated Relation type (sdpi:Relation) as the metric Relation type is anonymously specified and hence cannot be used anywhere except for AbstractMetricDescriptor. Conceptually, we're doing the same thing as we did with metrics. There has been quite some discussions and regrets as to why we did not add the Relation type to AbstractDescriptor - which would have spared us the SDPi version of a relation...

[PaulMartinsen]

@PaulMartinsen , there are two requirements defined by this PR (R0613 and R0614). Do they look fine in terms of RI "tagging"?

With the many changes you introduced lately (mostly with PR·469) I am often unsure if we are missing to use any important part of the new RI "tagging" functionality...

Yep. Looks okay to me.

[JavierEspina]

SDPi call - 19 Dec 2025: concluded that if the only blocker today is that Stefan did not have a chance to review in depth, we should approve and merge. If Stefan finds an issue later, we can address it then.