Definition of "Max Number of Devices" of the NW Profile

[tlohmar]

Problem description
There is no clear definition of the Max Number of Devices parameter within the NW Profile. The NW Profile defines the connectivity for the (reserved) networks. The key parameters are aggregated UL/DL Throughput, the QoS Profiles (one or more) and the default QOS profile.

The concept is that the aggregated UL/DL throughput defines the outer limits, e.g. 10Mbps UL and 5Mbps DL. The API invoker can combine the listed QoS profiles, as long as this is within the limits of the aggregated UL/DL throughput.

The main purpose of the Max Number of Devices parameter is to ensure, that enough NF resources are available. There is a significant difference, whether 10 or 10Mio devices are used within the reserved network.

In principle, there are two ways to define Max Number of Devices parameter:

1. Ball Park: the hard numbers are the aggregated UL/DL throughput. There may be more devices assigned to the reserved network than Max Number of Devices (but potentially not too many).
2. Hard Limit: the API does not allow any larger number of device accesses than Max Number of Devices.

Possible evolution

Define the meaning of the Max Number of Devices parameter of the NW profile.

Alternative solution

Additional context

[jordijoangimenez]

Following on this comment, also spotted here https://hub.5g-mag.com/Tech/pages/Network_APIs/CAMARA_DedicatedNetworks.html#5g-mags-self-assessment

To be discussed whether:

• There should be control on the maximum number of devices which are elegible to access (when assigned) the network resources, as you do not want to overbook resources which later on are not effectively used.
• There should be control on the maximum number of devices which can be assigned concurrently to the network resources, as you do not want devices to exceed the resources available.
• There should be control on the maximum number of devices which can invoke the API, as you want to have flexibility in accessing and releasing resources during operation (e.g. for a backup device).

In principle, number of maximum devices could:

• Be restricted for the effective concurrent assignment
• Be limited by a boundary so resources are not overbooked
• Be unlimited when it comes to invoking APIs and assigning/releasing different devices (as soon as aggregated number of devices concurrently using the resources is not exceeded)

[tlohmar]

Some additional considerations:

• Devices do not necessarily produce / consume traffic, i.e. not application packets flowing. For example, backup devices. Thus, the number of assigned devices is not necessarily an indication of the utilization of the reservation.
• The API invoker should have an interest to operate its devices & applications within the range of the reservation in order to ensure smooth operations of the assigned devices.
• Even when the API Provider allows a higher number of assigned devices than Max Number of Devices, the API provider may still not allow too many devices
  • Thus, it can make sense to define an Error Code for the case, when too many devices are assigned
  • It can be up for API provider implementation how many more devices than Max Number of Devices are allowed.