Improve adoption of Catalog Servers

[rafaelmag110]

WHAT AND WHY

The documentation provided upstream on the management domains concept explains very well its purpose as well as possible topologies to adopt.

In the mentioned docs, the usage of catalog servers seems to be the way forward to enable the correct adoption of the concept. It's of the interest of tractusx-edc to decide on the best approach, define the necessary implementation steps and enhance the documentation artifacts that ease the usage of these catalog servers.

Since there is already a primer that explains what might be the best approach to use catalog servers, the aim is to focus on what questions this document still doesn't answer regarding how the catalog servers should be leveraged, and take necessary actions towards answering those questions.

In my opinion, the documentation still misses the following points:

1. How is the catalog server distributed in tractusx? It seems it was included as an extension of the control plane runtime for the initial concept implementation, but the image in the documentation suggests it should be operated in its own management domain.
2. It could be a lack of understanding on my behalf but the usage of the Federated Catalog component is still confusing. It provides dataspace scraping functionalities, but can it also act as a catalog server?
3. The docs state the consumer EDC needs to take some steps before using the catalog server and the first one seems to be the creation of CatalogAssets. But the docs also state CatalogAssets are created by the provider. This is misleading. Maybe provide a pratical example of how a web of CatalogAssets should be created and queried?
4. One benefit of management domains is to enable a provider to control which entries of the returned DSP catalog are shown to a specific client by levering access policy and client credentials. An example of such usage could be provided.
5. The documention also shows an example of how to configure the current file-based implementation of the TargetNodeDirectory, but it states it should be used only for testing-purposes. A production grade implementation should be created.

HOW

• Create a decision record to define how a catalog server should be distributed in tractusx
• Create a decision record to define how tractusx-edc should implement the FC TargetNodeDirectory
• Implement the decisions taken
• Improve the documentation

[rafaelmag110]

@bmg13