Once the UE has downloaded the binary, the time when the UE performs the upgrade is left for the UE implementation, with possibly UE implementations seeking user input. In order to avoid impact to critical operations which would be executed during the UE unavailable period without prior knowledge from the core network and/or application function, this solution proposes that the UE reports its unavailable period to the network. The network takes the UE unavailable period into account when determining recommended time window for data transfer of the UE.
It is proposed that the NWDAF collects UE unavailable period from the UE via Data Collection AF. The application function negotiates recommended time window using background data transfer procedure. During this procedure, the PCF requests from the NWDAF the UE unavailable period. The PCF determines the recommended time for the application function to avoid the UE unavailable period.
In step 1, the AF invokes the Nnef_BDTPNegotiation_Create service and carries a "critical" indication. It indicates that this service is critical and the data transfer shall be performed avoid UE unavailable period.
In step 2, the NEF sends the "critical" indication within the Npcf_BDTPolicyControl_Create message to the H-PCF.
In step 5, the H-PCF interacts with the NWDAF using Nnwdaf_AnalyticsSubscription_Subscribe and requests for UE unavailable period for the corresponding UEs.
The NWDAF responds to the H-PCF with UE unavailable period using Nnwdaf_AnalyticsSubscription_Notify.
In step 3, the PCF requests UE unavailable period from the UDR.
In step 4, the UDR response to the H-PCF with UE unavailable period.
The H-PCF determines Background Data Transfer policies by taking into account the UE unavailable period.
The UE determines its unavailable period. The UE can determine its unavailable period based on UE implementation. The UE sends the UE unavailable period to the Data collection AF. This step reuses the data collection procedure in clause 184.108.40.206.1 of TS 23.288.
As described in key issue#1, UEs become unavailable (i.e. cannot interact with the 5G System) in the order of minutes whenever below events (hereafter called as events) are executed:
Silent reset at Modem;
Security patch updates;
Modem SW updates; and
Device reboot upon Modem setting changes via OMA-DM.
As UEs become unavailable without prior knowledge from the core network and/or application function while executing these events, it can impact critical operations of an application server if it depends on availability of the UE during the unavailability period (i.e. a period of time during which the UE is not available).
Thus this solution proposes a mechanism to co-ordinate the unavailability period between UE, 5GC and AF. So that UE can make informed decision on when to execute these events.
By re-using the procedures described in clause 220.127.116.11 of TS 23.502, the AF provisions the unavailability period parameter into the UDM/UDR, or the UDM/UDR can be configured with the unavailability period parameter based on operator policy.
The unavailability period parameter is provided to AMF by UDM using Nudm_SDM_Get response. Based on the unavailability period, AMF may choose to not initiate procedures which can force UE to enter deregistered state.
By re-using the procedures described in clause 18.104.22.168 of TS 23.502, AF provisions the unavailability period parameter into the PCF, or the PCF can be configured with the unavailability period parameter based on operator policy.
The unavailability period parameter is provided to AMF by PCF during AM policy association. Based on the unavailability period, AMF may choose to not initiate procedures which can force UE to enter deregistered state.
PCF uses the UE Configuration Update procedure for transparent UE Policy delivery as described in clause 22.214.171.124 of TS 23.502 to configure the unavailability period parameter into the UE.
The unavailability period parameter is the time slot which indicates the time in which UE is allowed to perform deregistration procedure to execute events described in clause 6.6.2. The unavailability period parameter can have multiple time slots for example 10:00-12:00 UTC, 14:00 - 18:00 UTC, 22:00-23:59 UTC etc. If UE is not able to execute the events for example due to the less storage capacity or insufficient battery level in the first time slot then UE should wait for the next time slot to execute the events.
The network takes unavailability period parameter into account to execute Device reboot upon Modem setting changes via OMA-DM procedures.
If UE have a user interface, how UE indicates the unavailability period to the user and take his consent to execute the events during unavailability period is outside the scope of this study as agreed in clause 2.
The unavailability period is configured by the HPLMN, based on agreements with the AF and operator policy, into the UE.