Should IP flows be supported in QoS Booking? #27
Replies: 8 comments
-
|
Maybe the API should allow the booking only down to the granularity of a device? And later, within the booked geography and time slot the API consumer can either request a provisioning for the (complete) device or for a specific flow (via quality-on-demand)? Then only the IP address within device parameter can be - optionally - used to identify the device at the time of the booking. Within later calls the API consumer might need to use the then current IP address to identify the same device. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for raising the issue. |
Beta Was this translation helpful? Give feedback.
-
|
@JoachimDahlgren is your question answered and we can close the issue. Alternatively we can convert to discussion. |
Beta Was this translation helpful? Give feedback.
-
|
My question was if IP flow granularity is to be supported by QoS Booking. The two answers kind of indicate that it should not be supported but that IP address could be used to identify the device at time of booking. If that is the case then this should be documented and devicePorts, applicationServer and applicationServerPorts ought to be removed not to lead people to think they can book for a specific flow. |
Beta Was this translation helpful? Give feedback.
-
@JoachimDahlgren My initial answer wasn't correct or at least misleading ... as @Masa8106 described in his answer it is possible for the network operator to derive all the necessary information which determines a flow:
|
Beta Was this translation helpful? Give feedback.
-
|
So I am to interpret your answers as "QoS Booking should support booking QoS for a future flow". That then leads me back to the original three questions
Note: It is probably a bad idea to use the IP address to do the booking as we could not respond back the MSISDN in the response. Thereby the only reference to the device in the booking (in case the IP address change) is an invalid IP address. |
Beta Was this translation helpful? Give feedback.
-
That is reopening all the discussions about the device object and which responses are allowed/possible. Yes, the IP address (which one: IPv4 or IPv6 ... IPv4 is normally not sufficient to identify a device) might be a last resort for specific use case. Preferable is the device identification within the access token (hint: there is no device identifier in the responses allowed). The newly defined "DeviceResponse" is only meant to indicate to the API Consumer which device property was used in case the API Consumer provided multiple ones. In short: in many cases there no device reference at all in the booking resource responses. |
Beta Was this translation helpful? Give feedback.
-
|
@JoachimDahlgren,
|
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
This issue is raised to trigger a discussion and to get further input. It originates from the work group meeting April 4th 2025 when this topic was first brought up.
Background
In the currently proposed first version of the QoS Booking API, it is possible to indicate which flow that sometime in the future is to get new QoS treatment. This flow is indicated by IP address and port of the device as well as of the application server.
With cellular connectivity, the IP address is assigned when the device connects to the network. There is no guaranteed that a disconnected device will get the same IP address when reconnected.
Questions
This raises the question if IP flow granularity is reasonable for the QoS booking API? Or should it rather be per device as is the case in the QoD Provisioning API?
In case we want it to be per flow:
Beta Was this translation helpful? Give feedback.
All reactions