issue_194_PR

[SyeddR]

What type of PR is this?

• PR for introducing application consumer Id as discussed in Add Application Function Id (afId) or Sponsor Id #194
  Add one of the following kinds:

• enhancement/feature

What this PR does / why we need it:

• Identifies the application consumer that initiates a qos session. The ID is systematically relayed to downstream systems, including but not limited to, the Network Exposure Function (NEF) and the Policy Control Function (PCF) within the 5G framework. The primary utility of this transmission is to facilitate the application of relevant policies, ensuring that the application consumer is appropriately tagged within usage records.

Which issue(s) this PR fixes:

Fixes #194

Special notes for reviewers:

For lack of the better term i am proposing "applicationConsumerId" as string type field. Open for suggestions

Changelog input

 release-note

Additional documentation

This section can be blank.

docs

[eric-murray]

Hi @SyeddR

I had understood that the applicationConsumerId must be associated with one, and only one application. Is that the case? If so, my preference would be for this to be a sub-property of the applicationServer property, the reason being that I'm not a fan of "flat" input parameter structures, and much prefer that related parameters be grouped, particularly when some (such as this one) are optional.

Of course, you could argue that this should also apply to applicationServerPorts, and I'd agree with you :-)

I also agree that adding this property to the ApplicationServer schema requires some thought, as we need to ensure that the API consumer provides at least one IP address, but there are ways in OAS that that can be mandated whilst still allowing applicationConsumerId to be optionally specified.

[SyeddR]

@eric-murray yes applicationConsumerId must be associated with one and only one application. I dont have a strong opinion on keeping it under applicationServer property or outside of it. Either way works for me.

[jlurien]

I suggested to open an issue on Identity and Consent WG about this topic, as identification of Application is being discussed there in the context of authentication, and access token should contain all required information.

[SyeddR]

As per the action items from the discussion in the last meeting.

1. Changed the naming from applicationConsumerId to applicationId to avoid confusion with the consumer of the application.
2. Added more description for the purpose of applicationId and how it is different than a clientId.
3. Moved the parameter under application server object.

[eric-murray]

Why the blank line here?

[SyeddR]

Fixed it.

[eric-murray]

@SyeddR
I'm fine with the proposal, but updated the ApplicationServer schema description to clarify that applicationId itself is not sufficient to identify the application server, and that an IP address is still required.

I'm a bit concerned that the schema definition will not catch API consumers who only provide the applicatonId (because that will satisfy the minProperties: 1 directive), so am wondering if we need to re-organise the schema so that a single ipAddress property is now required. One for discussion.

There appear to be some conflicts now with the main branch (which maybe I introduced - apologies if that is the case), Can you synchronise your fork with main to see if that fixes things?

[SyeddR]

There appear to be some conflicts now with the main branch (which maybe I introduced - apologies if that is the case), Can you synchronise your fork with main to see if that fixes things?

@eric-murray My fork shows all synced with latest updates. Can you try to resolve the conflict at your end.

[eric-murray]

@SyeddR
OK, fixed and now approved from my side. It seems my original edits were not properly merged into your own fork, but subsequent edits were, so all good now.

[RandyLevensalor]

Could we apply some constraints to this?

[RandyLevensalor]

Could we some constraints on the string size and values?
minLength: 3
maxLength: 256 (default value)
pattern: (should we exclude strings like "??? or ***"

[RandyLevensalor]

@SyeddR any response to this?

[hdamker]

Converted to draft status as agreed within our call on Nov 3rd.

[hdamker]

This PR is stale and can't be merged into the current (splitted) API structure anymore.