WO2022248032A1

WO2022248032A1 - Network node for qos notification control

Info

Publication number: WO2022248032A1
Application number: PCT/EP2021/064019
Authority: WO
Inventors: Antonio Consoli; Ali HAMIDIAN; Siva VAKEESAR
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2021-05-26
Filing date: 2021-05-26
Publication date: 2022-12-01
Also published as: CN117280666A

Abstract

The present disclosure relates to enhanced quality of service (QoS) notification control enabling a client device or an application function (AF) to be notified, in case of a QoS change to a QoS flow, whether the QoS change has been previously predicted or not. A session management function (SMF) or a policy control function (PCF) determines a prediction outcome for a QoS flow of a PDU session based on a comparison of a potential QoS change for the QoS flow with an actual change for the QoS flow. The prediction outcome may be a Boolean flag set to "true" if the potential QoS change for the QoS flow is the same as the actual change for the QoS flow and "false" in the opposite case. The SMF or PCF sends the determined prediction outcome for the QoS flow to a client device or an AF associated with the PDU session. Thereby, if an application knows about a QoS change in advance the application can e.g. mitigate the effects by putting in place measures. Furthermore, the present disclosure also relates to corresponding methods and a computer program.

Description

NETWORK NODE FOR QOS NOTIFICATION CONTROL

Technical Field

The present disclosure relates to a first network node and a second network node for enhanced QoS notification control and for QoS sustainability analytics enhancement. Furthermore, the present disclosure also relates to corresponding methods and a computer program.

Background

Vertical applications depend on the fact that service level specification (SLS) is fulfilled. The GSMA generic network slice template (GST) is used as the service level agreement (SLA) information for the communication between the vertical industry and the communication service provider. The SLA requirements need to be fulfilled from management plane and control plane perspectives in a coordinated way. The SLS includes ServiceProfile data type, The ServiceProfile data type represents the properties of network slice related requirement that should be supported by the network slice instance (NSI) in a 5G network. The network slice can be tailored based on the specific requirements and should adhere to SLA agreed between network slice customer (NSC) and network slice provider (NSP).

Deviation of the 5GS performance in any of the parameters described in the SLS may result in severe consequences to the application and money loss for the network operator, as a result of any penalty applied according to the SLA. SLS monitoring and measurement are the key components to ensure and verify the delivery of agreed SLS. Hence, the 5G system should provide monitoring and event management across the infrastructure, effectively and efficiently manage infrastructure 24 hours a days, 7 days a week (so-called 24x7), to ensure the shortest downtime and to provide immediate notification of performance degradation or service unavailability.

Fulfilment of SLS is based on a set of end-to-end key performance indicators (KPIs) that should be monitored by 5GS. If any deviation is observed on those parameters, the application should be notified and slice-level modification procedures should be triggered, to compensate for the deviation.

Summary

An objective of embodiments of the present disclosure is to provide a solution which mitigates or solves the drawbacks and problems of conventional solutions. The above and further objectives are solved by the subject matter of the independent claims. Further advantageous embodiments of the present disclosure can be found in the dependent claims.

According to a first aspect of the present disclosure, the above mentioned and other objectives are achieved with a first network node for a communication system, the first network node being configured to transmit a first message to a second network node, the first message indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node; receive a second message from the second network node, the second message indicating the potential QoS change for the QoS flow; determine a prediction outcome for the QoS flow based on a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow; and transmit a third message to a client device or an Application Function, AF, associated with the PDU session, the third message indicating the prediction outcome for the QoS flow.

A potential QoS change can herein e.g. be understood to mean a notification provided by a network node to a network consumer when the network consumer requests QoS sustainability analytics for an analytics target period.

An advantage of the first network node according to the first aspect is that the first network node can compare the notification of a QoS change that has happened with the predictions of QoS changes that it has received before on potential QoS change that may have happened. In this case the first network node knows if a QoS change that happened is something that was predicted in the past (in such case it also knows the prediction reference) and therefore differentiate from a systemic (or recurrent) issue from transient issues.

The normal operation of an application requires that the needed QoS is fulfilled by the 3GPP application. In general, any QoS change that is not triggered by the application may cause issues to the applications. Moreover, a QoS change that is predicted in advance is reducing potential harm for an application. In fact, if the application knows about a QoS change in advance it can mitigate its effects by putting in place countermeasures. Other benefits of the first network node includes:

• Support for 5GS Performance Predictability KPI as part of network slice SLS and SLS assurance procedure, which can be computed according to the prediction outcome and the relevant time periods when there is a QoS change. • Support for 5GS Performance Predictability KPI as part of the observed slice performance and application performance analytics.

• Ability for applications to provide requirements in terms of 5GS Performance Predictability in slice SLS.

• Ability for applications to receive information exposure on current 5GS Performance Predictability supported as part of the SLS.

In an implementation form of a first network node according to the first aspect, the first network node is configured to determine the prediction outcome for the QoS flow upon reception of a fourth message from a Radio Access Network, RAN, the fourth message indicating the actual change for the QoS flow.

An advantage with this implementation form is that by computing the prediction outcome upon reception of a fourth message from RAN, the first network node can assess if the 5GS is behaving in a predictable way according to the specific communication service, despite the QoS unfulfillment that is related to the actual QoS change. As explained above, predictability of the 5GS is a desired feature because a higher predictability determines a higher application availability. As the first network node is involved in session management of the specific communication service, an advantage in assessing the prediction outcome is that such information can be shared with the application as an indication of such predictability of the 5GS. The application can use such information to verify whether the 5GS is behaving in terms of predictability in conformance to the SLA. Another advantage is that the first network node may share such information with both ends of the application, to the U E-side of the application (e.g. via the SMF, AMF, RAN and UE) or the application side that is connected to the AF (e.g. via the PCF and the AF).

In an implementation form of a first network node according to the first aspect, the first network node is configured to transmit the first message to the second network node upon reception of a fifth message from the client device or the AF, the fifth message indicating a request for a prediction outcome of the QoS flow.

An advantage with this implementation form is that the first network node, involved in session management of the specific communication service, can combine the real time information of the session in question with the applicable predictions for the same session by transmitting the first message to the second network node upon reception of a fifth message from the client device or the AF. One example of such combination is the computation of the prediction outcome in case of a QoS change. The first network node may also use the prediction for the session in question also for other purposes, such as resource management, policy decisions or network configuration changes.

In an implementation form of a first network node according to the first aspect, the prediction outcome for the QoS flow is a Boolean flag.

An advantage with this implementation form is that a low overhead indication of the prediction outcome is provided.

In an implementation form of a first network node according to the first aspect, the second message further indicates an identifier of the potential QoS change for the QoS flow, and wherein the first network node is configured to identify the potential QoS change for the QoS flow based on the identifier.

An advantage with this implementation form is that the consumer of the NWDAF QoS Sustainability Analytics can use the novel introduced identifier refer to every prediction contained in a potential QoS change notification of analytics notification when relating it to an actual QoS change. In fact, the notification for cl. 6.9 of TS 23.288 (notification on potential QoS change or QoS Sustainability Analytics) may refer to a relatively long Analytics Target period, or a relatively large area (cell or tracking area) and therefore contain many predictions. The identifier can be used as a prediction reference in order to identify each predicted change of the KPI in question with respect to the preconfigured threshold defined by the analytics consumer.

In an implementation form of a first network node according to the first aspect, the first network node is configured to transmit a sixth message to the second network node, the sixth message indicating the prediction outcome for the QoS flow.

An advantage with this implementation form is that the first network node being involved in the session management of a specific communication service, the second network node can receive such information in real time whenever a QoS change happens to the communication service in question. The second network node may use such information to track and build reports on which communication services behave as predicted or not, according to communication service ID, time interval, serving nodes, serving network slice, etc. In an implementation form of a first network node according to the first aspect, the first network node is a Session Management Function, SMF, ora Policy Control Function, PCF, and wherein at least one of: the first message is a Nnwdaf_AnalyticsSubscription_Subscribe or

Nnwdaf_Analyticslnfo_Request; the second message is a Nnwdaf_AnalyticsSubscription_Notify or

Nnwdaf_Analyticslnfo_Response; the third message is a Npcf_SMPolicyControl_Update request or a PDU SESSION MODIFICATION COMMAND when the first network node is a SMF and the third message is a Npcf_PolicyAuthorization_Notify ora Npcf_SMPolicyControl_Update response when the first network node is a PCF; the fourth message is a Nsmf_PDUSession_UpdateSMContext request when the first network node is a SMF or a Npcf_SMPolicyControl_Update request when the first network node is a PCF; the fifth message is a Nsmf_PDUSession_UpdateSMContext Request or a Nsmf_PDUSession_CreateSMContext Request or a Npcf_PolicyAuthorization_Subscribe when the first network node is a SMF and the fifth message is a SM Policy Association Establishment or a Modification or a Npcf_PolicyAuthorization_Subscribe when the first network node is a PCF; and the sixth message is a Nsmf_EventExposure_Subscribe Notify.

An advantage with this implementation form is that the first network node can either be deployed in the SMF or PCF nodes part of the communication service session management and policy control.

According to a second aspect of the present disclosure, the above mentioned and other objectives are achieved with a second network node for a communication system, the second network node being configured to receive a first message from a first network node, the first message indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node; and transmit a second message to the first network node, the second message indicating the potential QoS change for the QoS flow and an identifier of the potential QoS change for the QoS flow. An advantage of the second network node according to the second aspect is that the consumer of the NWDAF QoS Sustainability Analytics can refer to every prediction contained in a potential QoS change notification of analytics notification when relating it to an actual QoS change. In fact the notification for cl. 6.9 of TS 23.288 (notification on potential QoS change or QoS Sustainability Analytics) may refer to a relatively long Analytics Target period, or a relatively large area (cell or tracking area) and therefore contain many predictions. The prediction reference may identify each predicted change of the KPI in question with respect to the preconfigured threshold defined by the analytics consumer.

In an implementation form of a second network node according to the second aspect, the second network node is configured to transmit a seventh message to the first network node, the seventh message indicating a request for a prediction outcome for the QoS flow; and receive a sixth message from the first network node, the sixth message indicating the prediction outcome for the QoS flow, the prediction outcome being a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow.

An advantage with this implementation form is that the prediction outcome can be used to compare the actual behaviour of the communication service with its predicted behaviour.

In an implementation form of a second network node according to the second aspect, the second network node is configured to determine a Key Performance Indicator, KPI, for the QoS flow and an observed time interval based on the prediction outcome for the QoS flow.

An advantage with this implementation form is that the KPI is an indication on how predictable is the 5GS behaving for the specific communication service or application in question. The KPI may be used as part of the SLA/SLS that a network slice customer requests from a network slice provider. 5GS services with higher values for such KPI may be charged at a higher price by the network slice provider because they may determine a higher application availability.

In an implementation form of a second network node according to the second aspect, the second network node is configured to determine the KPI for the QoS flow based on a ratio between: a sum of time intervals in which there is a QoS change for the QoS flow and in which a prediction outcome for the QoS flow is true, and a sum of time intervals in which there is a QoS change. An advantage with this implementation form is that this form of the KPI may be used to assess the potential gain or loss in terms of application availability. In fact, for the time intervals in which there is a QoS change and the prediction outcome is true, it is expected that the application availability can be higher with respect to the time intervals in which there is a QoS change and the prediction outcome is false. In fact, if the prediction outcome is true, the 5GS is behaving in a predictable way despite the QoS change and the application can be informed in advance about the upcoming QoS change. Such in-advance notification can be used by the application to counter-act for the upcoming QoS change and continue operation (increased availability).

In an implementation form of a second network node according to the second aspect, the second network node is configured to transmit the KPI to a NF consumer upon reception of a request for the KPI from the NF consumer.

An advantage with this implementation form is that if the NF consumer is a network slice management system, such KPI may be used to trigger potential slice adaptation procedures.

According to a third aspect of the present disclosure, the above mentioned and other objectives are achieved with a method for a first network node, the method comprises transmitting a first message to a second network node, the first message indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node; receiving a second message from the second network node, the second message indicating the potential QoS change for the QoS flow; determining a prediction outcome for the QoS flow based on a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow; and transmitting a third message to a client device or an AF associated with the PDU session, the third message indicating the prediction outcome for the QoS flow.

The method according to the third aspect can be extended into implementation forms corresponding to the implementation forms of the first network node according to the first aspect. Hence, an implementation form of the method comprises the feature(s) of the corresponding implementation form of the first network node.

The advantages of the methods according to the third aspect are the same as those for the corresponding implementation forms of the first network node according to the first aspect. According to a fourth aspect of the present disclosure, the above mentioned and other objectives are achieved with a method for a second network node, the method comprises receiving a first message from a first network node, the first message indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node; and transmitting a second message to the first network node, the second message indicating the potential QoS change for the QoS flow and an identifier of the potential QoS change for the QoS flow.

The method according to the fourth aspect can be extended into implementation forms corresponding to the implementation forms of the second network node according to the second aspect. Hence, an implementation form of the method comprises the feature(s) of the corresponding implementation form of the second network node.

The advantages of the methods according to the fourth aspect are the same as those for the corresponding implementation forms of the second network node according to the second aspect.

The present disclosure also relates to a computer program, characterized in program code, which when run by at least one processor causes said at least one processor to execute any method according to embodiments of the present disclosure. Further, the present disclosure also relates to a computer program product comprising a computer readable medium and said mentioned computer program, wherein said computer program is included in the computer readable medium, and comprises of one or more from the group: ROM (Read-Only Memory), PROM (Programmable ROM), EPROM (Erasable PROM), Flash memory, EEPROM (Electrically EPROM) and hard disk drive.

Further applications and advantages of the embodiments of the present disclosure will be apparent from the following detailed description.

Brief Description of the Drawings

The appended drawings are intended to clarify and explain different embodiments of the present disclosure, in which:

- Fig. 1 shows a first network node according to an embodiment of the present disclosure; - Fig. 2 shows a method for a first network node according to an embodiment of the present disclosure;

- Fig. 3 shows a second network node according to an embodiment of the present disclosure;

- Fig. 4 shows a method for a second network node according to an embodiment of the present disclosure;

- Fig. 5 shows a QoS notification procedure according to an embodiment of the present disclosure;

- Fig. 6 shows a slice QoE analytics procedure according to an embodiment of the present disclosure;

- Fig. 7 shows a QoS notification procedure according to an embodiment of the present disclosure;

- Fig. 8 shows a QoS sustainability analytics procedure according to an embodiment of the present disclosure;

- Fig. 9 shows a SLS assurance procedure according to an embodiment of the present disclosure;

- Fig. 10 shows a slice QoE analytics procedure according to an embodiment of the present disclosure;

- Fig. 11 shows SMF event notification according to an embodiment of the present disclosure;

- Fig. 12 shows a flow chart for QoS sustainability analytics subscription triggered by PDU session events according to an embodiment of the present disclosure;

- Fig. 13 shows a flow chart for QoS sustainability analytics subscription triggered by QoS notification control according to an embodiment of the present disclosure; and

- Fig. 14 shows calculation of an “asPredicted” Boolean flag according to an embodiment of the present disclosure.

Detailed Description

Some applications are adaptive by nature. This means that they may have several operation modes (which is also referred to application modes, e.g., normal mode, safe mode, and adapted mode) that can be triggered to function according to the QoS that can be provided by the network. The switch between application modes may not be immediate: it can last from a few hundreds of milliseconds up to several seconds to reconfigure a process control network according to a new set of parameters, and up to even several minutes to change relevant application level parameters. The transition between application modes must be completed before the QoS of the network is changed. Therefore, it becomes important that upcoming QoS changes are predicted ahead and notified to the application, so that the application can transition to an appropriate mode allowing continued operation under the upcoming QoS.

In this context, it becomes important that the SLS is also complemented by additional KPIs, as described by GSMA NEST: Performance prediction - availability; and Performance prediction - prediction frequency. GSMA GST v4.0 already includes performance prediction related E2E KPIs in the SLS template. These parameters are very important for industrial applications because performance prediction triggers the switch from normal to adapted mode (either way is possible too), which in turns helps decreasing application unavailability. Since decreasing application unavailability depends on the fact that the performance predictions are reliable (i.e., predictions are accurate enough to predict achievable 5GS performance) it is also important to introduce a new KPI that measures the quality of the performance prediction.

It is herein defined such KPI as follows: Performance prediction - 5GS performance predictability is an indication of the quality of the 5GS performance prediction service and therefore an indication related to how it can help reducing application unavailability. The present disclosure proposes to introduce this new E2E KPI in the 5GS and the way to calculate it. It is also introduced additional Information Elements (lEs) in the notifications generated by 5GS in case of performance degradations. The additional Information Elements (lEs) describe whether such performance degradation had been previously predicted or not. Such information is relevant for an application because it may still possible to handle performance degradations as long as those had been previously predicted, while unpredicted performance degradations may potentially generate more harm for the application. Some aspects of 5G performance prediction were introduced in 3GPP Rel-16 as specified in clause 6.9 of as a new set of analytics. Such analytics is called QoS Sustainability Analytics, allowing a service consumer such as an application or a 5GS network function (NF) to request predictions to be made for a target period (a time interval). Therefore, the present disclosure uses a terminology such as “performance prediction” to refer to the analytics service provided by such functionality. However, if, in future, new performance prediction services are introduced to predict 5GS performance, this new KPI can also refer to such new prediction services.

Furthermore, the present disclosure also focuses on calculating predictability “a posteriori” based on system observation and relevant collected analytics. Calculating predictability “a posteriori” means that first, the system is observed for a period and after sufficient information has been collected, the E2E KPI is calculated according to a mathematical formula. 5GS performance predictability can be defined by borrowing a few concepts from weather forecast science. In weather forecast science, given a current state of the atmosphere, an analog means the most similar scenario when compared with a repository of other states of the atmosphere in the past. The observed divergence in time of analogs (i.e., similar observed atmospheric states) provides an estimate of forecast divergence. In the 5GS, predictability can be defined by determining the time interval during which the system behaved as predicted (as in “analogs”), and the time intervals during which the system behaved differently than predicted.

Since different applications depend on prediction of different 5GS KPIs, 5GS predictability is relevant to the specific application that is under consideration. This means that the KPI should measure how good 5GS is in predicting those KPIs, which are important for that specific application, because the application may be sensitive to those KPIs. For example, motion control may depend on prediction of latency, because failure to comply on latency requirement can cause application unavailability, while tele-operated driving depends on prediction of latency as well as the data rate of the uplink path.

Therefore, the following is the proposed definition for 5GS performance predictability: 5GS performance predictability is - for a specific application - the ratio of the performance degradation or service unavailability events that occurred and had been notified in advance and the total performance degradation or service unavailability events, which occurred.

5GS performance predictability is an E2E KPI that measures how good the performance prediction service provided by 5GS to an application has been over the observed time window. Values of 5GS performance predictability close to 1 are an index of a good prediction service. The value of the 5GS performance predictability is directly connected with the application availability: the higher the predictability, the lower the application unavailability would be, since a higher portion of performance degradations may be compensated by adapted mode triggering instead of turning the system into safe mode.

Embodiments of the present disclosure are to define the 5GS performance predictability as a novel 5GS E2E KPI with the purpose of being part of the SLS of a network slice and the support of such new KPI in the 3GPP System. Embodiments of the present disclosure is to introduce support for new KPI with a solution which is based on the following existing mechanisms:

• Notification control mechanism, which is triggered by RAN in case of fulfilment/unfulfillment of a GBR QoS flow.

• QoS sustainability analytics, which supports prediction for potential QoS changes.

• Network slice observed service experience analytics, which supports analytics on service for a network slice. Since the notification control mechanism is supported for GBR type QoS flows, the bellow described embodiments in the 3GPP context covers, e.g., 5G system (5GS) performance predictability for GBR type QoS flows. The advantage of these embodiments of the present disclosure is that minimal changes are required on to the 5GS, as all of the services are already in place.

Embodiments of the present disclosure can therefore be related to at least two main concerns.

Concern 1 : the definition of a reporting and assurance procedure for such 5GS E2E KPI. Therefore, the same parameter shall be supported according to the following aspects of the 3GPP System: o In the defined end-to-end 5GS KPIs. o In the related NWDAF analytics reported as part of slice QoE and in the procedure of NSI and NSSI performance assurance o In the procedure of the service experience for an application.

Concern 2: the definition of mechanisms for exposing such KPI to the application, which also includes the definition of mechanisms to notify an application - in case of an event of performance degradation: o Which of those performance degradation/service unavailability events had actually been predicted or not for the application as a service consumer of the prediction analytics. o Which of those events may have been handled as adapted mode or not. o Which of those events caused application unavailability and which of them did not because unavailability of the application could be mitigated through adapted mode, which allows continued operation under specified circumstances.

Solution to concern 1 may involve modification of the QoS notification control procedure, e.g.:

• Modified procedure informing the UE that QoS for a specific QoS flow has been modified with the inclusion of an “asPredicted” Boolean flag upon UE request, as an additional information element (IE). The “asPredicted” Boolean flag corresponds to the previously described Boolean flag of the prediction outcome PO.

• Modified procedure for informing AF that QoS for a specific QoS flow has been modified with the inclusion of the “asPredicted” Boolean flag as an additional IE, upon request. In this notification, 5GS can also add the reference of the prediction as a further additional IE, which can inform the AF about when and how the QoS change was predicted. The reference of the prediction can be a new IE, which is described in the last bullet.

• New procedure for SMF-NWDAF interaction so that the SMF can become a consumer of the QoS sustainability analytics in order to determine the “asPredicted” Boolean flag. SMF can perform this new functionality by comparing - for each QoS flow of GBR type - the QoS that is currently enforced with the QoS that was previously predicted.

• Additional information elements on NWDAF for QoS sustainability analytics (optional): a unique event-id for each met or exceeded threshold, which can be used as “reference of the prediction”. The unique event-id corresponds to the previously described identifier of the potential QoS change.

Solution to concern 2 may involve support for 5GS performance predictability assurance, e.g.:

• SMF to include additional “asPredicted” Boolean flag in the events reported to the NWDAF.

• New triggers for the event-id= “QoS flow identifier (QFI) allocation” associated with RAN fulfilment/ unfulfillment of GBR QoS flows.

• Modified slice QoE analytics to include also the 5GS performance predictability to be reported as part of the slice QoE, to support appropriate assurance procedure.

Fig. 1 shows a first network node 100 according to an embodiment of the present disclosure. In the embodiment shown in Fig. 1 , the first network node 100 comprises a processor 102, a transceiver 104 and a memory 106. The processor 102 is coupled to the transceiver 104 and the memory 106 by communication means 108 known in the art. The first network node 100 may be configured for both wireless and wired communications in wireless and wired communication systems, respectively. The wireless communication capability may be provided with an antenna or antenna array 110 coupled to the transceiver 104, while the wired communication capability may be provided with a wired communication interface 112 coupled to the transceiver 104.

The processor 102 may be referred to as one or more general-purpose CPU, one or more digital signal processor (DSP), one or more application-specific integrated circuit (ASIC), one or more field programmable gate array (FPGA), one or more programmable logic device, one or more discrete gate, one or more transistor logic device, one or more discrete hardware component, one or more chipset. The memory 106 may be a read-only memory, a random access memory, ora non-volatile random access memory (NVRAM). The transceiver 104 may be a transceiver circuit, a power controller, an antenna, or an interface which communicates with other modules or devices. In embodiments, the transceiver 104 may be a separate chipset, or it is integrated with processor in one chipset. While in some implementations, the transceiver 104 the memory 106 and the processor 102 are integrated in one chipset.

That the first network node 100 is configured to perform certain actions can in this present disclosure be understood to mean that the first network node 100 comprises suitable means, such as e.g. the processor 102 and the transceiver 104, configured to perform said actions.

According to embodiments of the present disclosure the first network node 100 is configured to transmit a first message 510 to a second network node 300, the first message 510 indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node 100. The first network node 100 is configured to receive a second message 520 from the second network node 300, the second message 520 indicating the potential QoS change for the QoS flow. The first network node 100 is configured to determine a prediction outcome PO for the QoS flow based on a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow. The first network node 100 is further configured to transmit a third message 530 to a client device 610 or an AF 620 associated with the PDU session, the third message 530 indicating the prediction outcome PO for the QoS flow.

Fig. 2 shows a flow chart of a corresponding method 200 which may be executed in a first network node 100, such as the one shown in Fig. 1. The method 200 comprises transmitting 202 a first message 510 to a second network node 300, the first message 510 indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node 100. The method 200 comprises receiving 204 a second message 520 from the second network node 300, the second message 520 indicating the potential QoS change for the QoS flow. The method 200 comprises determining 206 a prediction outcome PO for the QoS flow based on a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow. The method 200 comprises transmitting 208 a third message 530 to a client device 610 or an AF 620 associated with the PDU session, the third message 530 indicating the prediction outcome PO for the QoS flow.

Fig. 3 shows a second network node 300 according to an embodiment of the present disclosure. In the embodiment shown in Fig. 3, the second network node 300 comprises a processor 302, a transceiver 304 and a memory 306. The processor 302 is coupled to the transceiver 304 and the memory 306 by communication means 308 known in the art. The second network node 300 may be configured for both wireless and wired communications in wireless and wired communication systems, respectively. The wireless communication capability may be provided with an antenna or antenna array 310 coupled to the transceiver 304, while the wired communication capability may be provided with a wired communication interface 312 coupled to the transceiver 304.

The processor 302 may be referred to as one or more general-purpose CPU, one or more digital signal processor (DSP), one or more application-specific integrated circuit (ASIC), one or more field programmable gate array (FPGA), one or more programmable logic device, one or more discrete gate, one or more transistor logic device, one or more discrete hardware component, one or more chipset. The memory 306 may be a read-only memory, a random access memory, ora non-volatile random access memory (NVRAM). The transceiver 304 may be a transceiver circuit, a power controller, an antenna, or an interface which communicates with other modules or devices. In embodiments, the transceiver 304 may be a separate chipset, or it is integrated with processor in one chipset. While in some implementations, the transceiver 304, the memory 306 and the processor 302 are integrated in one chipset. That the second network node 300 is configured to perform certain actions can in this present disclosure be understood to mean that the second network node 300 comprises suitable means, such as e.g. the processor 302 and the transceiver 304, configured to perform said actions.

According to embodiments of the present disclosure the second network node 300 is configured to receive a first message 510 from a first network node 100, the first message 510 indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node 100. The second network node 300 is further configured to transmit a second message 520 to the first network node 100, the second message 520 indicating the potential QoS change for the QoS flow and an identifier of the potential QoS change for the QoS flow.

Fig. 4 shows a flow chart of a corresponding method 400 which may be executed in a second network node 300, such as the one shown in Fig. 3. The method 400 comprises receiving 402 a first message 510 from a first network node 100, the first message 510 indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node 100. The method 400 comprises transmitting 404 a second message 520 to the first network node 100, the second message 520 indicating the potential QoS change for the QoS flow and an identifier of the potential QoS change for the QoS flow.

Fig. 5 shows signalling for a QoS notification control procedure according to an embodiment of the present disclosure. The QoS notification control procedure is performed between the first network node 100, the second network node 300, and a client device 610 or an AF 620. In embodiments, the first network node 100 is a session management function (SMF) or a policy control function (PCF) and the second network node 300 is a network data analytics function (NWDAF).

In step 1 in Fig. 5, the first network node 100 transmits a first message 510 to the second network node 300. The first message 510 indicates a request for a potential QoS change for a QoS flow of a PDU session established by the first network node 100. The PDU session may be a PDU session between the client device 610 and the AF 620.

The first network node 100 transmits the first message 510 to subscribe to notifications from the second network node 300 regarding potential QoS changes which are relevant for the QoS flow of the PDU session established by the first network node 100. When the first network node 100 is a SMF or PCF, the first message 510 may be a Nnwdaf_AnalyticsSubscription _Subscribe or Nnwdaf_Analyticslnfo_Request.

In a 3GPP implementation, the first message 510 may be seen as a request to receive predicted analytics for QoS sustainability analytics applicable for the QoS flow, e.g. the predicted analytics which are as relevant as possible specifically for the QoS flow in question. The first network node 100 may issue the subscription request for analytics, setting “Analytics ID” to "QoS Sustainability" and setting the “analytics filter information” containing the QoS requirements.

In embodiments, the first network node 100 transmit the first message 510 to the second network node 300 upon reception of a fifth message 550 from the client device 610 or the AF 620. The fifth message 550 indicates a request for a prediction outcome PO of the QoS flow. Hence, the first network node 100 may in embodiments subscribe to notifications for the QoS flow when triggered by the client device 610 or the AF 620. In the embodiment shown in Fig. 5, a fifth message 550 from the client device 610 or the AF 620 is indicated in step 1a.

In general, the client device 610 or the AF 620 requests the prediction outcome to provide such info to the application, so that the application knows if the specific QoS change had been previously predicted or not. In this respect, the second message 520 may further comprise an additional prediction reference IE which contains the identifier of the potential QoS change of the QoS flow. A QoS change that had been predicted in advance is reducing potential harmthan a QoS change that was not predicted as application can implement countermeasures. The predictability can be exposed also in the SLA and be related to charging for the 5GS service. A more predictable service is better than a less predictable one. The fifth message 550 may be transmitted as part of a PDU session establishment or PDU modification procedure. For example, when the client device 610 establish a PDU Session with at least one QoS flow of resource type guaranteed bit rate (GBR) or delay-critical GBR, when the client device 610 performs a PDU session modification request, or when the AF 620 performs an AF session request for the QoS of the QoS flow in question. By transmitting the fifth message 550 and hence the request for a prediction outcome PO of the QoS flow, the client device 610 or the AF 620 indicates that for the QoS flow and in case of an unfulfillment of the QoS, the client device 610 or the AF 620 wants to be notified whether the QoS change is according to a previous prediction or not.

When the first network node 100 is a SMF, the fifth message 550 may be a Nsmf_PDUSession_UpdateSMContext Request or a Nsmf_PDUSession_CreateSMContext Request or a Npcf_PolicyAuthorization_Subscribe. When the first network node 100 is a PCF, the fifth message 550 may be a SM Policy Association Establishment or a Modification or a Npcf_PolicyAuthorization_Subscribe.

The first message 510 transmitted from the first network node 100 in step 1 in Fig. 5 is received by the second network node 300. Hence, the second network node 300 obtains the request indicated in the first message 510, i.e. the request for a potential QoS change for a QoS flow of a PDU session established by the first network node 100. The second network node 300 is thereby informed that the first network node 100 wants to receive notifications related to potential QoS changes which are relevant for the QoS flow in question.

In step 2 in Fig. 5, the second network node 300 transmits a second message 520 to the first network node 100. The second message 520 indicates the potential QoS change for the QoS flow and may further in embodiments indicate an identifier (ID) of the potential QoS change event for the QoS flow. The second network node 300 may transmit the second message 520 in response to an application request. For example, the AF may send a Npcf_PolicyAuthorization_Subscribe because it wants to know any events related to a specific AF Session bound to a PDU Session since the AF acts as a control plane function connected to the application. The application is normally interested to know if something changes in the AF Session.

The second message 520 may indicate more than one potential QoS change events, e.g. different potential QoS change events which individually refer to different time intervals and/or different locations. Each potential QoS change event referring to a time interval and/or each location having a different identifier. For example: a potential QoS change message that refers to different time intervals may contain a potential QoS change event ID. For example: “potential QoS change event ID 1”, for the interval from 10.00 to 10.10 and QoS parameter X= 10; and “potential QoS change event ID 2”, for the time interval from 10.10 to 10.20 and QoS parameter X=9; and so on.

The first network node 100 receives the second message 520 indicating the potential QoS change for the QoS flow from the second network node 300. When the first network node 100 is a SMF or a PCF, the second message 520 may be a Nnwdaf_AnalyticsSubscription_Notify or Nnwdaf_Analyticslnfo_Response.

In embodiments where the second message 520 further indicates an identifier of the potential QoS change for the QoS flow, the first network node 100 can identify the potential QoS change for the QoS flow based on the identifier. The identifier may be used as evidence that in the past there had been a QoS prediction for the relevant prediction outcome. In this way whoever is receiving this prediction outcome knows it.

In step 3 in Fig. 5, the first network node 100 determines a prediction outcome PO for the QoS flow based on a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow. The first network node 100 may obtain the actual change for the QoS flow from a radio access network (RAN) 810 in a fourth message 540, as indicated in the optional step 3a in Fig. 5. Hence, the first network node 100 may in embodiments determine the prediction outcome PO for the QoS flow upon reception of a fourth message 540 from a RAN 810, where the fourth message 540 indicates the actual change for the QoS flow. Hence, the fourth message 540 function as a trigger in this case. The RAN 810 may transmit the fourth message 540 when the QoS of the QoS flow is changed, e.g. due to that the RAN 810 can no longer fulfil the previous QoS of the QoS flow. The fourth message 540 may be transmitted from the RAN 810 to the first network node 100 via an AMF, if the first network node 100 is a SMF, and via an AMF and a SMF, if the first network node 100 is a PCF. When the first network node 100 is a SMF, the fourth message 540 may be a Nsmf_PDUSession_UpdateSMContext request. When the first network node 100 is a PCF, the fourth message 540 may be a Npcf_SMPolicyControl_Update request.

The prediction outcome PO for the QoS flow may in embodiments be a Boolean flag indicating whether the potential QoS change is the same as the actual change or not, i.e. whether the QoS change was as predicted or not. For example, the prediction outcome PO may be determined to be set to a first value e.g., “true”/”1” if the potential QoS change is the same as the actual change for the QoS flow and determined to be set to a second value e.g., “falseTO” if the potential QoS change is different than the actual change for the QoS flow.

In step 4 in Fig. 5, the first network node 100 transmits a third message 530 to the client device 610 or the AF 620 associated with the PDU session. The third message 530 indicates the determined prediction outcome PO for the QoS flow. When the first network node 100 is a SMF, the third message 530 may be a Npcf_SMPolicyControl_Update request or a PDU SESSION MODIFICATION COMMAND. When the first network node 100 is a PCF, the third message 530 may be a Npcf_PolicyAuthorization_Notify or a Npcf_SMPolicyControl_Update response.

When the prediction outcome PO for the QoS flow is a Boolean flag, the Boolean flag may be set to “true”/”1” if the potential QoS change is the same as the actual change for the QoS flow and set to “falseTO” if the potential QoS change is different from the actual change for the QoS flow. In case the Boolean flag is set to “trueTT, the third message 530 may further indicate an identifier of the potential QoS change for the QoS flow which can be used by the client device 610 or the AF 620 to identify the potential QoS change.

Fig. 6 shows signalling for a slice quality of experience (QoE) analytics procedure according to an embodiment of the present disclosure. The slice QoE analytics procedure is performed between the first network node 100, the second network node 300, and a network function (NF) consumer 710. In embodiments, the first network node 100 is a SMF or a PCF and the second network node 300 is a NWDAF.

In step 1 in Fig. 6, the second network node 300 transmit a seventh message 570 to the first network node 100, the seventh message 570 indicting a request for a prediction outcome PO for the QoS flow. In this way, the second network node 300 can indicate to the first network node 100 that the second network node 300 wants to know, in case of a QoS change, whether the potential QoS change is the same as the actual change or not. The seventh message 560 may be a Nsmf_EventExposure_Subscribe.

In step 2 in Fig. 6, the first network node 100 transmits a sixth message 560 to the second network node 300, the sixth message 560 indicating the prediction outcome PO for the QoS flow. When the first network node 100 is a SMF or a PCF, the sixth message 560 may be a Nsmf_EventExposure_Subscribe Notify. The first network node 100 may transmit the sixth message 560 upon determining a prediction outcome PO for the QoS flow based on an indication from the RAN 810 that the QoS of the QoS flow has changed, as described with reference to Fig. 5.

The second network node 300 receives the sixth message 560 from the first network node 100 and hence the indicated prediction outcome PO for the QoS flow. The prediction outcome PO may be a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow. When the prediction outcome PO for the QoS flow is a Boolean flag, the Boolean flag may be set to “true” if the potential QoS change is the same as the actual change for the QoS flow and set to “false” if the potential QoS change is different from the actual change for the QoS flow. In case the Boolean flag is set to “true” the sixth message may further indicate an identifier of the potential QoS change for the QoS flow which can be used to identify the potential QoS change.

In step 3 in Fig. 6, the second network node 300 determines a key performance indicator (KPI) for the QoS flow and an observed time interval based on the prediction outcome PO for the QoS flow. The KPI indicates the QoE associated with the QoS flow. The second network node 300 may determine the KPI for the QoS flow based on a ratio between: a sum of time intervals in which there is a QoS change for the QoS flow and in which a prediction outcome PO for the QoS flow is true, and a sum of time intervals in which there is a QoS change. In other words, the determined KPI for the QoS flow indicates the ratio of the performance degradation or service unavailability events that occurred and were predicted/notified in advance and the total performance degradation or service unavailability events which occurred. The KPI for the QoS flow may hence be seen as a measure of how good the performance prediction service has been over an observed time window, where a value close to 1 indicates a good prediction service. The value of KPI for the QoS flow is directly connected with the application unavailability. The higher the value of KPI for the QoS flow, the lower the application unavailability, since a higher portion of performance degradations may be compensated by adapted mode triggering instead of turning the system into safe mode.

The second network node 300 may share the determined KPI with one or more NF consumers. In the embodiment shown in Fig. 6, the second network node 300 transmits the determined KPI to a NF consumer 710 in step 4. In embodiments, the second network node 300 transmits the KPI to the NF consumer 710 upon reception of a request for the KPI from the NF consumer 710, as indicated in optional step 1a in Fig. 6. The NF consumer 710 may request the KPI from the second network node 300 using a Nnwdaf_Analyticslnfo_Request or a Nnwdaf_AnalyticsSubscription_Subscribe. In the following disclosure yet further embodiments of the present disclosure will be presented for providing deeper understanding of the disclosed solution. In this respect, also the following embodiments are set in a 3GPP context hence the terminology, expressions, protocols, interfaces, and architecture used. Therefore, a client device corresponds to a UE in such implementation examples. However, embodiments of the present disclosure are not limited thereto and can be implemented in any suitable communication system. Generally, the implementation of embodiments of the present disclosure may include but is not limited to modifications to the following 3GPP system procedures:

1. QoS notification control procedure, including UE and AF notification, and the case of alternative QoS profile.

2. NWDAF QoS sustainability procedure.

3. SLS assurance procedure.

4. NWDAF slice QoE analytics procedure.

5. SMF event notification for event “QFI allocation”.

The following sections describe details of the proposed modifications, as well as implementation examples of embodiments of the present disclosure.

Changes to QoS notification control procedure including UE and AF notification Fig. 7 shows signalling related to the modifications proposed to the existing QoS notification control procedure. Purpose for introducing the novel aspects according to embodiments of the present disclosure is the following:

• To enable 5GS to notify, in case of an arising QoS change to any of the QoS flows of GBR or delay-critical GBR type, both the AF and the UE if such QoS change had been previously predicted (in a previous analytics notification) or not, i.e. the prediction outcome. The novel information may be added in an “asPredicted” Boolean flag IE, which can assume the values “true” if the QoS change had been previously predicted, or “false” in the opposite case. “Had been previously predicted,” means that the application has received a QoS sustainability analytics notification that is applicable for the specific QoS change.

• In case the QoS change had been predicted, also to include in the notification the relevant reference of such prediction, i.e. a prediction reference.

The additional information may be added as novel information to the notifications that are sent to the UE and/or the AF for GBR and delay-critical GBR QoS flows. The additional novel information may be sent when the AF or the UE request such information. When performing the request to the 5GS, the AF or the UE may set a specific “asPredicted” requested Boolean flag, to indicate the request of such additional novel information. The procedure will now be described with reference to Fig. 7. The procedure assumes as preconditions the following steps:

In step 1 in Fig. 7, the UE has established the PDU session with at least one QoS flow of resource type GBR or delay-critical GBR. As a novel functionality, the UE has set for a GBR or delay-critical GBR type QoS flow the “asPredicted” requested Boolean flag to the value “true”, to indicate that - for the QoS flow and in case of an unfulfillment of the QoS - the UE wants to be notified whether the QoS change is according to a previous prediction or not. The ’’asPredicted” requested Boolean flag may also be set in the context of a PDU session modification request performed by the UE, or as part of an AF session request with QoS issued by an AF for the QoS of the QoS flow in question. The “asPredicted” requested Boolean flag may be set when SMF has knowledge that the related QoS flow exists and has an assigned QoS profile. The establishment of the QoS flow may result from several operations. Therefore, the “asPredicted” requested Boolean flag might be set in any operation that causes the establishment of the QoS flow.

In step Ibis in Fig. 7, after receiving the request for establishment or modification of the PDU session with the relevant QoS flow with the “asPredicted” requested Boolean flag toggled to “true”/”1”, the SMF subscribes to the NWDAF to receive the predicted analytics for QoS sustainability analytics applicable for the QoS flow. The SMF may issue the subscription request to receive predicted analytics, which are as relevant as possible specifically for the QoS flow in question. The SMF may further issue the subscription request for analytics, e.g., by setting “Analytics ID” to "QoS Sustainability" and setting the “analytics filter information” containing at least one of the QoS requirements as follows:

• In the “5QI” filter, it will set the 5QI of the QoS flow in question, including applicable additional QoS parameters and the corresponding values, e.g. guaranteed flow bit rate (GFBR), maximum flow bit rate (MFBR), etc.

• The QoS characteristics attributes including resource type, packet delay budget (PDB), packet error rate (PER), and their values are set according to the relevant values of the QoS flow in question.

• The time window for the “analytics target period” is set for a window that starts at current time and lasts for an appropriate time interval in the future for which the QoS flow is expected to operate or last. This interval is application dependent and is set either according to a static configuration or according to previously collected analytics, e.g. 5 min, 10 min, 1 hour, etc.

• The location information, as provided by the AMF, is set according to the current location and the area extension according to where the QoS flow is expected to be operated in the near future. The extension of such area is application dependent and may be decided according to a static configuration or according to previously collected analytics, e.g. one cell or more cells.

• Reporting threshold(s) is set for the QoS flow retainability KPI, in order to trigger an NWDAF notification any time the QoS is predicted to be different from the QoS that is currently enforced for the QoS flow in question.

• Single network slice selection assistance information (S-NSSAI) is set according to the S- NSSAI of the PDU Session. In case the PDU session references more than one S-NSSAIs it has to consider the S-NSSAIs where the QoS flow is expected to operate in the near future.

In step 2 in Fig. 7, the AF has subscribed to the PCF for the events of QoS fulfilment and/or unfulfillment. This is done when the AF issues a NpcfPolicyAuthorization_Subscribe for the QoS flow in question. This request may also include the “asPredicted” requested Boolean flag set to “true” to indicate that the AF has requested the additional information if the QoS change is according to a previous prediction or different from what had been previously predicted.

The following steps describe the events and the actions that occur in the network in case of a QoS fulfilment or unfulfillment event for the QoS flow of GBR or delay-critical GBR type is triggered by next generation RAN (NG-RAN). It is assumed that for such QoS flow, notification control has been activated and the “asPredicted” requested Boolean flag is set to “true”, as described in steps 1-2 in Fig. 7.

In step 3 in Fig. 7, a QoS change happens because NG-RAN is not able to fulfil a GBR or delay-critical GBR of the QoS flow. Therefore NG-RAN issues a notification control message towards the core network (CN), a PDU SESSION RESOURCE NOTIFY (i.e. PDU session ID and N2 SM information) containing PDU session resource notify transfer IE (i.e. QoS flow notify list IE, notification cause IE, and current QoS parameters set index IE) for the QoS flow in question indicating the new QoS that NG-RAN is able to fulfil. This message is received by the AMF.

In step 4 in Fig. 7, the AMF upon reception of the message from the NG-RAN issues an Nsmf_PDUSession_UpdateSMContext request (i.e. SM Context ID and N2 SM information) towards the SMF that is in charge of managing the PDU session in question. The message may be used by the AMF to inform the SMF about the QoS notification control notification and contains the information for the relevant QoS flow, including the new QoS that can be fulfilled for the QoS flow. In case the QoS flow has a set of alternative QoS profiles, it also contains the IE currentQosProfilelndex = N or nullQoSProfilelndex = true, where N is the QoS profile that NG-RAN can fulfil or if no alternative QoS profile can be fulfilled, nullQoSProfilelndex = true.

In step 5 in Fig. 7, triggered by the message received in step 4, the SMF informs the PCF about the new QoS applied for the QoS flow in question by issuing a

Npcf_SMPolicyControl_Update request with the additional novel information (i.e. asPredicted) which the SMF can determine according to the notification received following message in step Ibis:

• “asPredicted” Boolean flag information which is set to “true” if the QoS is according to what had been previously predicted or “false” if the QoS is different from what was predicted, or there is no prediction that can be associated with the relevant QoS change event that has been reported by message received in step 4.

• “prediction Reference” information containing the unique id of the predicted event.

The novel information is added for every unfulfillment event. The SMF can add such information inside the qncReports array containing one or more elements of

QosNotificationControllnfo. Inside such element, the notifType item of QosNotifType may take the value either of GUARANTEED or of NOT_GUARANTEED. Since the novel information is added for the unfulfillment events, it is added for the notifType with value “NOT_GUARANTEED”.

In step 6 in Fig. 7, if the AF has requested notification as described in step 2, the PCF informs the AF issuing a Npcf_PolicyAuthorization_Notify containing the new QoS for the QoS flow in question as well as the novel predictability information as specified in the step 5 in Fig. 7.

In step 7 in Fig. 7, the PCF sends an Npcf_SMPolicyControl_Update response to the SMF following the request in step 5.

In step 8 in Fig. 7, as one of the events triggered by the message received in step 4, the SMF issues a PDU SESSION MODIFICATION COMMAND to the UE with the authorized QoS rules information. If “asPredicted” requested Boolean flag is set to “true” for the PDU session, the SMF also includes the novel information of the optional “asPredicted” Boolean flag. Such novel IE is set to “true” in case the QoS change included in the authorized QoS rules of the PDU SESSION MODIFICATION COMMAND is according to a previous delivered prediction, or to “false” in case there is no reference of a prediction that is related to the QoS change in question. In step 9 in Fig. 7, the UE sends PDU SESSION MODIFICATION COMPLETE message to the SMF in response to the PDU SESSION MODIFICATION COMMAND message received in step 8 and indicates an acceptance of the PDU SESSION MODIFICATION COMMAND message. As an alternative, the UE may send a PDU SESSION MODIFICATION REJECT if the UE rejects the PDU SESSION MODIFICATION COMMAND. There is no novel IE or modifications in the reject message.

In step 10 in Fig. 7, the SMF responds to the message received in step 4 by the AMF by issuing an Nsmf_PDUSession_UpdateSMContext response confirming the update of the PDU session with relation to the changes to the QoS flow in question.

Changes to NWDAF QoS sustainability procedure

According to an embodiment of the present disclosure changes to the QoS sustainability analytics are herein also provided. The QoS sustainability analytics is used to compute predictions on the QoS in a determined location area and time window. Such changes introduce a unique ID for each predicted event of QoS change, with the purpose of using such an id as reference when there is a notification of a QoS change that happened, every time the AF or UE requests such information.

Fig. 8 shows a QoS sustainability analytics procedure according to an embodiment of the present disclosure.

In step 1 in Fig. 8, the NF acting as NF consumer, e.g. the SMF as described in Fig. 7, issues a subscription for QoS sustainability analytics towards the NWDAF, providing relevant analytics filter information and a relevant analytics target period. Therefore, the NF consumer transmits a Nnwdaf_Analyticslnfo_Request or a Nnwdaf_AnalyticsSubscription_Subscribe message to the NWDAF.

In step 2 in Fig. 8, the NWDAF receives the subscription request and starts computing the analytics response. To do so, the NWDAF receives user plane function (UPF) information and in step 3 receives information by data collection from the operations, administration and maintenance (OAM).

In step 4 in Fig. 8, the NWDAF derives the requested analytics based on the collected data from the OAM. As part of the analytics information that is provided to the NF consumer, the NWDAF introduces a novel unique ID for each of the events that are part of the prediction. This unique ID is called “Prediction ID”. The NWDAF outputs a number of QoS sustainability analytics entries which is limited by the maximum number of objects provided as part of analytics reporting information. Each entry shall have a unique prediction ID. The NWDAF can detect the need for notification (a potential QoS change notification), or in such case adding an entry to the notification, based on comparing the requested analytics of the target 5QI against the reporting threshold(s) provided by consumer in any cell over the requested analytics target period. Table 2 below describes lEs of the QoS sustainability analytics including the novel “Prediction id”.

Table 2

For each analytic in the list included in the notification, each of the crossed reporting thresholds, a unique ID is introduced that can be used to later refer to the prediction, e.g. in a later QoS change notification. Providing information about when and how the QoS change was predicted. Finally, in step 5 in Fig. 8, the NWDAF responds to the NF consumer by transmitting a Nnwdaf_Analyticslnfo_Response or a Nnwdaf_AnalyticsSubscription_Notify message to the NF consumer.

Changes to SLS assurance procedure According to an embodiment of the present disclosure, the SLS assurance procedure is enhanced to include 5GS performance predictability. The SLS assurance procedure is used for performance assurance for network slice instance(s) (NSI(s)) or network slice subnet instance(s) (NSSI(s)) and is shown in Fig. 9. In step 1 in Fig. 9, the authorized NSMS_Consumer requests NSMS_Producer to allocate a new NSI. The NSMS_Producer consumes provisioning services provided by NSSMS_Producer to create the NSI. After NSI allocation, the NSMS_Producer and NSSMS_Producer perform NSI and NSSI performance supervision.

After step 1 a loop is executed involving steps 2 and 3 with optional steps 4a and 4b which be described in the following disclosure.

In step 2 in Fig. 9, the NSMS_Producer or NSSMS_Producer may get slice QoE analytics provided by a NWDAF. In addition, 5GS performance predictability is added as relevant metric related to the performance of QoS sustainability analytics (as NWDAF service) as the optimization of such analytics together with the use of prediction to act appropriate application countermeasure can help decreasing application unavailability.

In step 3 in Fig. 9, the NSMS_Produceror NSSMS_Producer checks whether the performance requirements can be met by NSI or NSSI by utilizing the end-to-end KPIs, performance measurements and slice QoE analytics provided by the NWDAF.

In step 4a in Fig. 9, if the performance requirements of NSI cannot be met, the NSMS_Producer triggers the NSI modification procedure. The NSMS_Producer modifies the capacity of the network slice or modifies the network slice configuration to guarantee the performance requirements.

In step 4b in Fig. 9, if the performance requirements of NSSI cannot be met, the NSSMS_Producer of the CN modifies virtualized resources and the configuration of 5G core network (5GC) NFs to guarantee the performance requirements. NSSMS_Producer of the access network reconfigures RRMPolicy to optimize performance.

This SLS assurance procedure covers slice-level performance parameters on which the NWDAF reports the fulfilment of the measured slice-level KPIs, see Fig. 10 for details, and operations support systems (OSS) adjusts the NSI parameters any time there is a deviation from promised SLS on any observed parameter. As an example, parameters such as allocated radio resources, virtual network function (VNF) instances, central processor unit (CPU) power, memory, etc can be adjusted any time service level is not “as required”. Changes to NWDAF slice QoE analytics procedure

Fig. 10 shows a NWDAF slice QoE analytics procedure according to an embodiment of the present disclosure. The NWDAF slice QoE analytics procedure is modified with the introduction of the following main changes:

• When requesting the event notification to SMF for the event id “QFI allocation”, the NWDAF may toggle the “asPredictedRequested” Boolean flag to “true” to indicate that in case of QoS change, the NWDAF is requesting also the information if the change is according to what was predicted or not. Such functionality requires a modification of the trigger for the event id “QFI allocation” that is described in the next bullet.

• The event ID “QFI allocation” may be triggered also for an additional event which is the event in which the QoS flow profile is changed because of a RAN fulfillment or unfulfillment event.

• The information included in the Nsmf_EventExposure_Notify may include together with the information on the QFI allocation the information if the QoS change event is according to a previous prediction or not, as well as the reference to such predicted event, i.e. the related unique id.

• When replying to the NF consumer, the NWDAF may compute the end-to-end KPI “5GS performance predictability” and include the new analytics in the Nnwdaf_Analyticslnfo_Request Response or in the Nnwdaf_AnalyticsSubscription_Notify.

In step 1 in Fig. 10, a NF consumer sends an analytics request/subscribe message to a NWDAF, where analytics ID = service experience, target of analytics reporting = any UE, analytics filter information = application ID, analytics target period S-NSSAI, DNN, area of interest, to the NWDAF by invoking a Nnwdaf_Analyticslnfo_Request or a Nnwdaf_AnalyticsSubscription_Subscribe. The NF consumer can be a NSSMS_Producer as depicted in Fig. 9.

In step 2a in Fig. 10, the NWDAF collects input data from relevant NFs by invoking Nnef_EventExposure_Subscribe (in case the NF is an AF) or Naf_EventExposure_Subscribe (in case of a generic 5GC NF). In the solution proposed one of the applicable NFs is the SMF and the service invoked is the Nsmf_EventExposure_Subscribe Request for the Event ID = QFI allocation. Assuming the SMF triggers the Event ID = QFI allocation also for NG-RAN fulfilment/unfulfillment events, which is every time SMF receives a QoS notification control for a GBR type QoS flow (novel modification introduced in this I PR), the NWDAF can add a specific novel information element in this Nsmf_EventExposure_Subscribe Request to signal that the prediction outcome (e.g. the asPredicted flag) is also requested by the SMF. In step 2b in Fig. 10, the SMF acknowledge the request by issuing a Nsmf_EventExposure_Subscribe Response to the NWDAF. When the notification control message is received by the SMF and the Event ID = QFI allocation is also triggered, the SMF can issue as in step 2c a Nsmf_EventExposure_Subscribe Notify also adding the “asPredicted” Boolean flag toggled to “true” if the QoS change is as predicted or there is an applicable prediction, or “false” if the QoS change is different than what was predicted in the Nsmf_EventExposure_Notify in step 2c. In case “asPredicted” is set to “true”, the SMF may also provide a reference to the prediction by incorporating in the notification/response also the unique prediction id.

Steps 3a to 3d in Fig. 10 are not changed in this disclosure compared to the present procedure in cl. 6.4.5 of TS 23.288.

In step 4 in Fig. 10, according to the proposed implementation the NWDAF derives e.g. by computing also the end-to-end 5GS requested predictability analytics for a network slice.

In step 5 in Fig. 10, the NWDAF provides the data analytics, i.e. the observed service experience which may be a range of values, to the NF consumer. The NWDAF provides the data analytics by means of either Nnwdaf_Analyticslnfo_Request or Nnwdaf_AnalyticsSubscription_Notify response to the NF consumer. The relevant method may be chosen depending on the service used in step 1, indicating how well the used QoS parameters satisfy the service mean opinion score (MoS) agreed between the mobile network operator (MNO) and the end user or between the MNO and the external application service provider (ASP).

If the NF consumer is a PCF and determines that the application SLA is not satisfied, the NF consumer may take into account the observed service experience and the operator policies including SLA and required service experience which can be a range of values to determine new QoS parameters to be applied for the service.

Changes to procedure related to SMF event notification for event QFI allocation According to an embodiment of the present disclosure the possibility to trigger the event ID “QFI allocation” also for NG-RAN fulfilment/unfulfillment events is introduced. In this case, the following changes are introduced to the procedure related to SMF event notification for event “QFI allocation”, shown in Fig. 11. In step 1 in Fig. 11, the SMF sends a event notification to a NF service consumer for a QFI allocation, including the following information: a) QFI of the allocated QoS flow id for the application as "qfi" attribute; b) Data network name (DNN) of the allocated PDU session as "dnn" attribute; c) Slice of the allocated PDU session as "snssai" attribute; and d) The description of the application traffic as "appld", "fDescs" or "ethfDescs" attribute.

The following novel information is proposed to be added to the message provided by the SMF to the NF service consumer, e.g. the NWDAF: e) The “asPredicted” Boolean flag set to “1” if the QFI allocation event is triggered in case of QoS fulfilment/unfulfillment and the QoS is as predicted or to “0” if the QFI allocation event is triggered in case of QoS fulfilment/unfulfillment and the QoS is different than what was predicted; and f) Reference “prediction ID” which references the relevant prediction ID as set by the NWDAF in the QoS sustainability analytics.

Adding this new information, together with triggering the event ID “QFI allocation” for every notification of QoS fulfilment/unfulfillment, introduces the possibility for the NWDAF to be notified of every QoS change from an initial QoS and every time the QoS change reverts back to the initial QoS for the QoS flow. This notification is provided with the information if the QoS change was predicted or not. According to such information the NWDAF may compute for the specific QoS flow the 5GS performance predictability KPI according to the formula:

In other words, the KPI for the QoS flow is determined based on a ratio between: a sum of time intervals in which there is a QoS change for the QoS flow and in which a prediction outcome for the QoS flow is true, and a sum of time intervals in which there is a QoS change. The NWDAF may calculate the KPI globally for the whole 5GS, or for a specific 5QI, or for a set of 5Qls. Further implementations

In further implementations of the present disclosure, most of the novel logic may be added to the SMF, as described with reference to Fig. 12, which shows a flow chart for QoS sustainability analytics subscription from the SMF triggered by the events PDU session establishment, modification, or release.

In step 1 , the SMF detects a PDU Session establishment, modification, or release event which causes a QoS flow of type GBR or delay-critical GBR to be allocated or modified. This triggers the SMF to check the following conditions in step 2-4 to determine whether to subscribe to QoS sustainability analytics:

• Whether the QoS flow has set the NotifControl=true in step 2.

• Whether the involved UE or AF have set the “asPredicted” requested Boolean flag to “true” in step 3.

• Whether as the result of the PDU session establishment or modification the QoS flow is allocated in step 4, i.e. whether the outcome of the check in step 4 is “Yes”.

If all the three conditions are true, the SMF node subscribes to QoS sustainability analytics in step 5, thereby subscribing to receive from the NWDAF the predicted analytics for the QoS flow in question. The SMF may subscribe to the QoS sustainability analytics with the following parameters 5QI = QFI, QoS characteristics attributes including resource type, PDB, PER and their values, location = characteristics of QoS flow, S-NSSAI = S-NSSAI of PDU session, [optional maximum number of objects], analytics target period = from now to the next 10-15 minutes, reporting threshold = based on the alternative QoS profile (if available) or based on the characteristics of the QoS flow e.g. MFBR.GFBR, notification correlation ID, notification target address, etc.

In case the SMF detects a PDU session establishment, modification or release event in step 1 , which causes a QoS flow of type GBR or delay-critical GBR to be released or deallocated, the SMF checks the following conditions in step 2-4 to determine whether to unsubscribe to QoS sustainability analytics:

• Whether the QoS flow has set the NotifControl=true in step 2.

• Whether as the result of the PDU session establishment or modification the QoS flow is released or deallocated in step 4, i.e. whether the outcome of the check in step 4 is “No”. If all the three conditions are true, the SMF node unsubscribes to QoS sustainability analytics in step 6. In this way, every time the QoS flow is deallocated, the subscription may be cancelled by the SMF.

The NWDAF detects the need for notification about a potential QoS change based on comparing the expected values for the KPI of the target 5QI against the reporting threshold(s) provided by the consumer in any cell in the requested area for the requested analytics target period. The expected KPI values are derived from the statistics for the 5QI obtained from OAM. OAM information may also include planned or unplanned outages detection and other information that is not in scope for 3GPP to discuss in detail. The analytics feedback contains the information on the location and the time when a potential QoS change may occur and what reporting threshold(s) may be crossed. The actions that the SMF performs when receiving a NWDAF notification are detailed in the next paragraph.

According to an embodiment of the present disclosure novel logic is introduced to the SMF allowing the NF to compute the value of the “asPredicted” Boolean flag every time a QoS notification control message is received by NG-RAN and notify the UE and/or the AF accordingly. Notification to the UE and/or the AF may happen if such novel information was requested, i.e. the “asPredicted” Boolean flag is set to “true” either for the UE or for the AF.

Fig. 13 shows QoS sustainability analytics subscription from the SMF triggered by QoS notification control according to an embodiment of the present disclosure. When a QoS notification control event is received from NG-RAN through the AM F in step 1 , the SMF checks for each of the QoSFIowNotifyltem contained in the QoSFowsNotifyList if they refer to a notification cause of unfulfillment in step 2. In case they refer to an unfulfillment event, i.e. the outcome of the check in step 2 is “Yes”, the SMF checks whether the notification contains a reference to an alternative QoS profile that is fulfilled in alternative to the original one in step 3, i.e. that the nullQoSProfilelndex = false. If the outcome of the check in step 3 is “Yes”, the SMF compares the QoS contained in the new enforced QoS profile with the QoS that has been predicted for the QoS in question in step 4, according to location, time interval and other information included in the subscription. If the outcome of the check in step 3 is “No”, the “asPredicted” Boolean flag is set to “false” in step 5.

In case the new enforced QoS profile matches with that of the prediction, i.e. the outcome of the check in step 4 is “Yes”, the “asPredicted” Boolean flag is set to “true” in step 6. In case no applicable prediction is found, i.e. the outcome of the check in step 4 is “No”, the asPredicted Boolean flag is set to “false” in step 5. Any time a new prediction is received for a managed QoS flow, the SMF may store the prediction internally as part of the local PDU context and save it for later in case a QoS unfulfillment event is received for one of the QoS flows for which “asPredicted” is requested.

Therefore, according to an embodiment of the present disclosure shown in Fig. 14, the SMF may use received information from the NWDAF on QoS Sustainability Analytics notification to internally store the predictions that are applicable for a specific QoS flow of GBR or delay- critical GBR type over an analytics target period in the future. Every time the SMF receives a QoS sustainability analytics notification, the SMF may determine if it applies for the QoS flow in question. If the prediction reports a predicted potential QoS change, in case a QoS notification control message is also received for QoS unfulfillment, the SMF may determine if the QoS change had been predicted or not. The SMF may then set the “asPredicted” Boolean flag to “true” in case it had been predicted or “false” in case it had not been predicted in the event id “QoS allocation” notification, or in the non-access stratum (NAS) notification to the UE or in the notification towards the PCF and AF. In case “asPredicted” is set to “true”, the prediction reference IE may be set according to the value of the prediction id contained in the applicable analytics entry in the notification for QoS sustainability analytics.

For subscription to SMF events, the SMF may trigger the event ID “QFI allocation” in the occurrence of a QoS notification control and add the “asPredicted” Boolean flag and “prediction reference” if applicable to the event notification provided to the NF consumer. In such implementation, the NF consumer may be the NWDAF.

An alternative implementation could be based on the definition of a new SMF event which could be triggered in the occurrence of a QoS notification control, containing the same lEs of the above implementation, i.e. the “asPredicted” Boolean flag and “prediction reference” if applicable. In such case, the NWDAF configures as the NF consumer.

The first network node 100 herein may be denoted as a SMF (MME in LTE and LTE-A) or a PCF (PCRF in LTE and LTE-A) in 5G. The second network node 300 herein may be denoted as a NWDAF in 5G NR. The NWDAF may be a generic function/node configured for communication in 3GPP related LTE and LTE-Advanced.

Furthermore, any method according to embodiments of the present disclosure may be implemented in a computer program, having code means, which when run by processing means causes the processing means to execute the steps of the method. The computer program is included in a computer readable medium of a computer program product. The computer readable medium may comprise essentially any memory, such as a ROM (Read- Only Memory), a PROM (Programmable Read-Only Memory), an EPROM (Erasable PROM), a Flash memory, an EEPROM (Electrically Erasable PROM), or a hard disk drive.

Moreover, it is realized by the skilled person that embodiments of the first network node 100 and the second network node 300 comprise the necessary communication capabilities in the form of e.g., functions, means, units, elements, etc., for performing the solution. Examples of other such means, units, elements and functions comprise: processors, memory, buffers, control logic, encoders, decoders, rate matchers, de-rate matchers, mapping units, multipliers, decision units, selecting units, switches, interleavers, de-interleavers, modulators, demodulators, inputs, outputs, antennas, amplifiers, receiver units, transmitter units, DSPs, MSDs, TCM encoder, TCM decoder, power supply units, power feeders, communication interfaces, communication protocols, etc. which are suitably arranged together for performing the solution.

Especially, the processor(s) of the first network node 100 and the second network node 300 may comprise, e.g., one or more instances of a Central Processing Unit (CPU), a processing unit, a processing circuit, a processor, an Application Specific Integrated Circuit (ASIC), a microprocessor, or other processing logic that may interpret and execute instructions. The expression “processor” may thus represent a processing circuitry comprising a plurality of processing circuits, such as, e.g., any, some or all of the ones mentioned above. The processing circuitry may further perform data processing functions for inputting, outputting, and processing of data comprising data buffering and device control functions, such as call processing control, user interface control, or the like.

Finally, it should be understood that the present disclosure is not limited to the embodiments described above, but also relates to and incorporates all embodiments within the scope of the appended independent claims.

Claims

1. A first network node (100) for a communication system (500), the first network node (100) being configured to transmit a first message (510) to a second network node (300), the first message (510) indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node (100); receive a second message (520) from the second network node (300), the second message (520) indicating the potential QoS change for the QoS flow; determine a prediction outcome (PO) for the QoS flow based on a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow; and transmit a third message (530) to a client device (610) or an Application Function, AF, (620) associated with the PDU session, the third message (530) indicating the prediction outcome (PO) for the QoS flow.

2. The first network node (100) according to claim 1, configured to determine the prediction outcome (PO) for the QoS flow upon reception of a fourth message (540) from a Radio Access Network, RAN, (810), the fourth message (540) indicating the actual change for the QoS flow.

3. The first network node (100) according to claim 1 or 2, configured to transmit the first message (510) to the second network node (300) upon reception of a fifth message (550) from the client device (610) or the AF (620), the fifth message (550) indicating a request for a prediction outcome (PO) of the QoS flow.

4. The first network node (100) according to any one of the preceding claims, wherein the prediction outcome (PO) for the QoS flow is a Boolean flag.

5. The first network node (100) according to any one of the preceding claims, wherein the second message (520) further indicates an identifier (ID) of the potential QoS change for the QoS flow, and wherein the first network node (100) is configured to identify the potential QoS change for the QoS flow based on the identifier (ID).

6. The first network node (100) according to any one of the preceding claims, configured to transmit a sixth message (560) to the second network node (300), the sixth message (560) indicating the prediction outcome (PO) for the QoS flow.

7. The first network node (100) according to any one of the preceding claims, wherein the first network node (100) is a Session Management Function, SMF, or a Policy Control Function, PCF, and wherein at least one of: the first message (510) is a Nnwdaf_AnalyticsSubscription_Subscribe or Nnwdaf_Analyticslnfo_Request; the second message (520) is a Nnwdaf_AnalyticsSubscription_Notify or Nnwdaf_Analyticslnfo_Response; the third message (530) is a Npcf_SMPolicyControl_Update request ora PDU SESSION MODIFICATION COMMAND when the first network node (100) is a SMF and the third message (530) is a Npcf_PolicyAuthorization_Notify or a Npcf_SMPolicyControl_Update response when the first network node (100) is a PCF; the fourth message (540) is a Nsmf_PDUSession_UpdateSMContext request when the first network node (100) is a SMF or a Npcf_SMPolicyControl_Update request when the first network node (100) is a PCF; the fifth message (550) is a Nsmf_PDUSession_UpdateSMContext Request or a Nsmf_PDUSession_CreateSMContext Request or a Npcf_PolicyAuthorization_Subscribe when the first network node (100) is a SMF and the fifth message (550) is a SM Policy Association Establishment or a Modification or a Npcf_PolicyAuthorization_Subscribe when the first network node (100) is a PCF; and the sixth message (560) is a Nsmf_EventExposure_Subscribe Notify.

8. A second network node (300) for a communication system (500), the second network node (300) being configured to receive a first message (510) from a first network node (100), the first message (510) indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node (100); and transmit a second message (520) to the first network node (100), the second message (520) indicating the potential QoS change for the QoS flow and an identifier (ID) of the potential QoS change for the QoS flow.

9. The second network node (300) according to claim 8, configured to transmit a seventh message (570) to the first network node, the seventh message (570) indicting a request for a prediction outcome (PO) for the QoS flow; and receive a sixth message (560) from the first network node (100), the sixth message (560) indicating the prediction outcome (PO) for the QoS flow, the prediction outcome (PO) being a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow.

10. The second network node (300) according to claim 9, configured to determine a Key Performance Indicator, KPI, for the QoS flow and an observed time interval based on the prediction outcome (PO) for the QoS flow.

11. The second network node (300) according to claim 10, configured to determine the KPI for the QoS flow based on a ratio between: a sum of time intervals in which there is a QoS change for the QoS flow and in which a prediction outcome (PO) for the QoS flow is true, and a sum of time intervals in which there is a QoS change.

12. The second network node (300) according to claim 10 or 11, configured to transmit the KPI to a NF consumer (710) upon reception of a request for the KPI from the NF consumer (710).

13. A method (200) for a first network node (100), the method (200) comprising transmitting (202) a first message (510) to a second network node (300), the first message (510) indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node (100); receiving (204) a second message (520) from the second network node (300), the second message (520) indicating the potential QoS change for the QoS flow; determining (206) a prediction outcome (PO) for the QoS flow based on a comparison of the potential QoS change for the QoS flow with an actual change for the QoS flow; and transmitting (208) a third message (530) to a client device (610) or an AF (620) associated with the PDU session, the third message (530) indicating the prediction outcome (PO) for the QoS flow.

14. A method (400) for a second network node (300), the method (400) comprising receiving (402) a first message (510) from a first network node (100), the first message (510) indicating a request for a potential QoS change for a QoS flow of a PDU session established by the first network node (100); and transmitting (404) a second message (520) to the first network node (100), the second message (520) indicating the potential QoS change for the QoS flow and an identifier (ID) of the potential QoS change for the QoS flow.

15. A computer program with a program code for performing a method according to any one of claims 13 to 14 when the computer program runs on a computer.