WO2024023555A1 - Managing communication network resources per user session based on user quality of experience (qoe) - Google Patents

Abstract

Embodiments include methods for managing communication network resources based on end-user quality of experience (QoE). Such methods include collecting performance-related data from a RAN and a core network comprising the communication network, and correlating at least the collected performance-related data into a plurality of records corresponding to a plurality of user data sessions with the communication network. Such methods include training one or more artificial intelligence/machine learning (AI/ML) models based on the correlated data records. Each AI/ML model maps between network quality-of-service (QoS) metrics and service-specific QoE metrics. Such methods include configuring communication network resources to carry one or more second user data sessions, based on the trained AI/ML model and on QoE requirements of respective services comprising the second user data sessions. Configuring communication network resources can include, for example, enforcing service-specific policies based on QoE requirements and/or traffic steering based on predicted QoE and QoE requirements.

Description

MANAGING COMMUNICATION NETWORK RESOURCES PER USER SESSION BASED ON USER QUALITY OF EXPERIENCE (QOE)

TECHNICAL FIELD

The present disclosure relates generally to communication networks, and more specifically to techniques for managing communication network resources based on end-user quality of experience (QoE) for individual user data sessions rather than network-level performance-related data such as quality of service (QoS) metrics.

BACKGROUND

Currently the fifth generation (“5G”) of cellular systems, also referred to as New Radio (NR), is being standardized within the Third-Generation Partnership Project (3GPP). NR is developed for maximum flexibility to support multiple and substantially different use cases. These include enhanced mobile broadband (eMBB), machine type communications (MTC), ultra-reliable low latency communications (URLLC), side-link device-to-device (D2D), and several other use cases.

Figure 1 shows a high-level view of an exemplary 5G network 100, including a Next Generation RAN (NG-RAN) 199 and a 5G Core (5GC) 198. NG-RAN 199 can include a set of gNodeB’s (gNBs) connected to the 5GC via one or more NG interfaces, such as gNBs 100, 150 connected via interfaces 102, 152, respectively. In addition, the gNBs can be connected to each other via one or more Xn interfaces, such as Xn interface 140 between gNBs 100 and 150. With respect the NR interface to UEs, each of the gNBs can support frequency division duplexing (FDD), time division duplexing (TDD), or a combination thereof.

NG-RAN 199 is layered into a Radio Network Layer (RNL) and a Transport Network Layer (TNL). The NG-RAN architecture, i.e., the NG-RAN logical nodes and interfaces between them, is defined as part of the RNL. For each NG-RAN interface (NG, Xn, Fl) the related TNL protocol and the functionality are specified. The TNL provides services for user plane transport and signaling transport.

The NG RAN logical nodes shown in Figure 1 include a central (or centralized) unit (CU or gNB-CU) and one or more distributed (or decentralized) units (DU or gNB-DU). For example, gNB 100 includes gNB-CU 110 and gNB-DUs 120 and 130. CUs (e.g., gNB-CU 110) are logical nodes that host higher-layer protocols and perform various gNB functions such controlling the operation of DUs. Each DU is a logical node that hosts lower-layer protocols and can include, depending on the functional split, various subsets of the gNB functions. As such, each of the CUs and DUs can include various circuitry needed to perform their respective functions, including processing circuitry, transceiver circuitry (e.g., for communication), and power supply circuitry. A gNB-CU connects to gNB-DUs over respective Fl logical interfaces, such as interfaces 122 and 132 shown in Figure 1. The gNB-CU and connected gNB-DUs are only visible to other gNBs and the 5GC as a gNB. In other words, the Fl interface is not visible beyond gNB-CU.

Centralized control plane protocols (e.g., PDCP-C and RRC) can be hosted in a different CU than centralized user plane protocols (e.g. , PDCP-U). For example, a gNB-CU can be divided logically into a CU-CP function (including RRC and PDCP for signaling radio bearers) and CU- UP function (including PDCP for UP). A single CU-CP can be associated with multiple CU-UPs in a gNB. The CU-CP and CU-UP communicate with each other using the El-AP protocol over the El interface, as specified in 3GPP TS 38.463 (vl5.4.0). Furthermore, the Fl interface between CU and DU (see Figure 1) is functionally split into Fl-C between DU and CU-CP and Fl-U between DU and CU-UP. Three deployment scenarios for the split gNB architecture shown in Figure 1 are CU-CP and CU-UP centralized, CU-CP distributed/CU-UP centralized, and CU-CP centralized/CU-UP distributed.

Open RAN (O-RAN) ALLIANCE is a community of mobile operators and RAN vendors working towards an open, intelligent, virtualized, operationally efficient, and fully interoperable RANs. To achieve these goals, the community has defined an O-RAN Architecture with key functions and interfaces. Various specifications published by O-RAN work groups (WGs). For example, O-RAN WG1 is concerned with use cases and overall architecture. One general principle is that O-RAN architecture and interface specifications shall be consistent with 3GPP architecture and interface specifications, to the extent possible.

Figure 2 shows a high-level view of the O-RAN architecture, including interfaces Al, 01, Open Fronthaul M-plane, and 02. These interfaces connect the Service Management and Orchestration (SMO) framework to O-RAN network functions (NFs) and the Open Cloud (O- Cloud). Additionally, there is an interface between SMO and external systems that can provide enrichment data. Also shown is the NG interface between O-RAN NFs and the NG-Core, which is consistent with the NG interface with 5GC shown in Figure 1.

The O-RAN Architecture Description defines the following three control loops with respective latencies:

• Real Time (RT) Control Loop (<10 ms);

• Near-RT RIC Control Loop (10-1000 ms); and

• Non-RT RIC Control Loop (>1000 ms).

Use cases for Non-RT RIC and Near-RT RIC control loops are fully defined by O-RAN, but O- RAN only defines relevant interactions with other O-RAN nodes or functions for the RT control loop (which performs radio scheduling, HARQ, beamforming, etc.). The Non-RT RIC provides the Al interface to the Near-RT RIC. One task of Non-RT RIC is to provide policy-based guidance, machine learning (ML) model management, and enrichment information to support intelligent RAN optimization by the Near-RT RIC (e.g., for radio resource management, RRM). The Non-RT RIC can also perform intelligent RRM in longer, non-RT intervals (e.g., greater than 1 second).

The Non-RT RIC can use data analytics and artificial intelligence (AI)/ML training and inference to determine RAN optimizations, for which it can leverage SMO services such as data collection from and provisioning to the O-RAN nodes. These actions are performed by Non-RT RIC Applications (rApps). The Non-RT RIC also includes the Non-RT RIC Framework, which is internal to the SMO Framework, logically terminates the Al interface, and exposes all required functionality and services to rApps.

Figure 3 shows a more detailed view of the O-RAN architecture shown in Figure 2. For example, this figure shows functions, units, and interfaces in the O-RAN architecture (e.g., O-DU, O-CU-CP, O-CU-UP, Xn, NG, Fl, El) that correspond to similarly named ones in NG-RAN and 5GC. Figure 3 also shows an additional E2 interface between the Near-RT RIC and various units or functions, as well as an open radio unit (O-RU) that communicates with the O-DU and an O- eNB that terminates the Uu interface for LTE.

O-Cloud shown in Figures 2-3 is a cloud computing platform of physical infrastructure nodes that host various O-RAN functions (e.g., Near-RT RIC, O-CU-CP, O-CU-UP, O-DU etc.), supporting software components (e.g., Operating System, Virtual Machine Monitor, Container Runtime, etc.), and appropriate management and orchestration functions.

The SMO framework is responsible for RAN domain management, and includes the following capabilities and/or functionality:

• FCAPS interface to O-RAN NFs such as Performance Management (PM), Configuration Management (CM), Fault Management (FM), File Management, Communications Surveillance (Heartbeat), Trace, etc.;

• Non-RT RIC for RAN optimization; and

• O-Cloud Management, Orchestration, and Workflow Management.

SMO does not define any formal interfaces to Non-RT RIC, such that SMO implementations may make their respective design choices for creating boundaries with the Non-RT RIC Framework, including no boundaries at all.

O-RAN WG1 has also published a use case analysis report (v5.0) and a corresponding detailed specification (v5.0) that identifies 18 different O-RAN use cases. Of particular interest are use case 4 (quality-of-experience, QoE, optimization) and use case 5 (traffic steering). SUMMARY

Importantly, these two use cases are user-centric rather than network-centric. For example, the O-RAN specifications require that traffic steering should be based on UE-specific conditions rather than more general traffic conditions in the cell in which the UE operates. However, conventional traffic steering solutions are based on cell-level key performance indicators (KPIs) and/or quality-of-service (QoS) metrics, and are generally independent of any specific application or service being used by a UE.

As another example, QoE relates to end-to-end (E2E) user experience for specific applications, including for interactive applications such as virtual reality (VR) or augmented reality (AR) that are traffic-intensive and latency-sensitive. Conventional techniques are based on a semi-static QoS framework that includes different predefined combinations of QoS characteristics such as packet delay, packet error rate, priority level, etc. However, these general QoS characteristics do not guarantee E2E QoE for specific services, specific network slices, or specific cells over various traffic loads and mixtures of traffic types.

Embodiments of the present disclosure provide specific improvements to E2E QoE and traffic steering in a RAN, such as by providing, enabling, and/or facilitating solutions to exemplary problems summarized above and described in more detail below.

Embodiments include methods (e.g., procedures) for managing communication network resources based on end-user QoE. These exemplary methods can be performed by a network management system for the communication network (e.g., SMO).

These exemplary methods can include collecting performance-related data from a RAN (e.g., an O-RAN) and a core network (CN) comprising the communication network. These exemplary methods can also include correlating at least the collected performance-related data into a plurality of records corresponding to a plurality of user data sessions with the communication network. These exemplary methods can also include training one or more AI/ML models based on the correlated data records, wherein each AI/ML model maps between network QoS metrics and service-specific QoE metrics. These exemplary methods can also include configuring communication network resources to carry one or more second user data sessions, based on the trained AI/ML model and on QoE requirements of a service comprising the second user data sessions.

In some embodiments, each AI/ML model is trained based on the correlated data records using the network QoS metrics as input features and the service-specific QoE data as output labels. In some variants, the service-specific QoE data includes one or more of the following: opinion score, service availability, bitrate statistics, playback errors, rebuffering, startup time, startup failures, and duration of user data session. In some embodiments, these exemplary methods can also include determining a plurality of service-specific policies for network resource management, based on the one or more trained AI/ML models and on QoE requirements of a respective plurality of services. Each servicespecific policy includes corresponding network QoS requirements. In some of these embodiments, configuring communication network resources to carry one or more second user data sessions in includes enforcing service-specific policies for the respective second user data sessions.

In some embodiments, configuring communication network resources to carry one or more second user data sessions can also include the following: using the one or more AI/ML models, predicting QoE for the respective services comprising the active user data sessions based on a detected change in network QoS metrics; and changing one or more network resources or settings for each second user data session based on the respective QoE requirements of the services comprising the second user data sessions. In particular, the one or more second user data sessions (whose network resources or settings are changed) comprise services having predicted QoE that do not meet corresponding QoE requirements. In some of these embodiments, the one or more network resources or settings that are changed include one or more of the following: serving cell, resource allocation within serving cell, serving radio access technology (RAT), serving RAN node, and QoS settings. These operations can also be referred to as “traffic steering”.

Other embodiments include network management systems (e.g., SMOs, cloud-based systems, etc.) configured to perform operations corresponding to any of the exemplary methods described herein. Other embodiments include non-transitory, computer-readable media storing program instructions that, when executed by processing circuitry, configure such network management systems to perform operations corresponding to any of the exemplary methods described herein.

These and other embodiments described herein can provide a currently missing connection between controllable system resources and end-user QoE: AI/ML models that can predict achieved QoE based on the dedicated resources and QoS settings, with respect to service classification and network conditions. As a result, network resource policies are based on user- or service-level QoE requirements rather than network QoS metrics. Also, embodiments base policy decisions on information provided by a service-specific, per-flow monitoring and analytics system instead of a static QoS framework (e.g., 5QI) used in conventional systems.

Thus, embodiments regulate QoE with better granularity without requiring overprovisioning of network resources to provide safety margins, as done in conventional QoS-based techniques. As such, embodiments can facilitate more efficient use of network resources and reduce network load. Additionally, embodiments can provide user-centric QoE optimization and traffic steering in contrast to conventional cell-centric solutions that are based on network resources and QoS without regard to actual user experience.

These and other objects, features, and advantages of embodiments of the present disclosure will become apparent upon reading the following Detailed Description in view of the Drawings briefly described below.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows a high-level views of an exemplary 5G/NR network architecture.

Figures 2-3 show different views of the O-RAN architecture.

Figure 4 shows a high-level block diagram of a system configured for session-based traffic steering with end-user QoE optimization, according to embodiments of the present disclosure.

Figure 5 is a more detailed block diagram that illustrates an exemplary configuration of the system shown in Figure 4 in the O-RAN architecture.

Figure 6 shows a more detailed diagram of a system configured for session-based traffic steering with end-user QoE optimization, according to embodiments of the present disclosure.

Figures 7-8 shows two implementation option for integrating a system according to embodiments of the present disclosure into an O-RAN architecture.

Figure 9 (which includes Figures 9A-B) shows a flow diagram of an exemplary method (e.g., procedure) for managing communication network resources based on end-user QoE, according to various embodiments of the present disclosure.

Figure 10 shows a communication system according to various embodiments of the present disclosure.

Figure 11 shows a network node according to various embodiments of the present disclosure.

Figure 12 shows host computing system according to various embodiments of the present disclosure.

Figure 13 is a block diagram of a virtualization environment in which functions implemented by some embodiments of the present disclosure may be virtualized.

DETAILED DESCRIPTION

Some of the embodiments contemplated herein will now be described more fully with reference to the accompanying drawings. Other embodiments, however, are contained within the scope of the subject matter disclosed herein, the disclosed subject matter should not be construed as limited to only the embodiments set forth herein; rather, these embodiments are provided as examples to convey the scope of the subject matter to those skilled in the art. Generally, all terms used herein are to be interpreted according to their ordinary meaning in the relevant technical field, unless a different meaning is clearly given and/or is implied from the context in which it is used. All references to a/an/the element, apparatus, component, means, step, etc. are to be interpreted openly as referring to at least one instance of the element, apparatus, component, means, step, etc., unless explicitly stated otherwise. The steps of any methods and/or procedures disclosed herein do not have to be performed in the exact order disclosed, unless a step is explicitly described as following or preceding another step and/or where it is implicit that a step must follow or precede another step. Any feature of any of the embodiments disclosed herein can be applied to any other embodiment, wherever appropriate. Likewise, any advantage of any of the embodiments can apply to any other embodiments, and vice versa. Other objects, features, and advantages of the enclosed embodiments will be apparent from the following description.

Note that the description given herein focuses on a 3GPP cellular communications system and, as such, 3GPP terminology or terminology similar to 3GPP terminology is oftentimes used. However, the concepts disclosed herein are not limited to a 3GPP system. Furthermore, although the term “cell” is used herein, it should be understood that (particularly with respect to 5G NR) beams may be used instead of cells and, as such, concepts described herein apply equally to both cells and beams.

As briefly mentioned above, O-RAN WG1 has also published a use case analysis report (v5.0) and a corresponding detailed specification (v5.0) that identifies 18 different O-RAN use cases. Of particular interest are use case 4 (quality-of-experience, QoE, optimization) and use case 5 (traffic steering).

Some native 5G applications such as virtual reality (VR) or augmented reality (AR) are traffic-intensive and latency-sensitive. Conventional techniques are based on a semi-static QoS framework that includes different predefined combinations of QoS characteristics such as packet delay, packet error rate, priority level, etc. In 5G, each of these combinations is identified by an index called 5G QoS Identifier (5QI). When a user initiates a specific service, the 5GC assigns that service to a QoS flow (between UE and 5GC, via RAN) that has QoS characteristics (and associated 5QI) that corresponds to the service’s QoS requirements.

However, these general QoS characteristics do not guarantee E2E QoE for specific services, specific network slices, or specific cells over various traffic loads, mixtures of traffic types, and/or radio conditions. O-RAN use case 4 is based on the expectation that applicationlevel QoE estimation/prediction can help deal with such uncertainty, improve radio resource efficiency, and ultimately improve user experience.

The main objective for use case 4 is to ensure QoE optimization be supported within the O-RAN architecture and its open interfaces. Multi-dimensional data (e.g., user traffic data, QoE measurements, network measurement report, etc.) can be acquired and processed via AI/ML models to support traffic recognition, QoE prediction, and QoS enforcement decisions. AI/ML models can be trained offline and but inference will be executed in real-time. O-RAN QoE optimization should be a general solution that supports QoE for any specific service such as Cloud VR, video, etc.

Conventionally, network operators have manually configured parameters for congested cells to improve QoE of users. In contrast, O-RAN is intended to facilitate QoE optimization in real-time with proactive closed-loop network optimization. Congested cells or cell resources should be detected in real-time and resources allocated dynamically to satisfy user QoE. In the O- RAN architecture, this is done by the Non-RT RIC, Near-RT RIC, and E2 Nodes (e.g., eNB, gNB) based on AI/ML models.

For example, one task of Non-RT RIC is constructing an AI/ML model and training it based on data retrieved from SMO and network level measurements. Trained models (which can be considered “policies”) are sent to Near-RT RIC to be used for managing RAN parameters. In particular, the Near-RT RIC will use trained AI/ML models to assist QoE optimization tasks such as predictions about application/traffic type classification, QoE, and available bandwidth. To support these operations, E2 nodes provide performance management (PM) information to SMO over the 01 interface. Also, the Non-RT RIC and E2 Nodes can communicate RRM behavior updates over E2 interface to support QoS enforcement.

Regarding O-RAN traffic steering use case 5, 5G systems will support many different combinations of radio access technologies (RATs) including LTE (licensed spectrum), NR (licensed or unlicensed), and Wi-Fi (unlicensed). The number of frequency bands being used in mobile networks has also grown tremendously. These variables make it challenging to switch or steer traffic in a balanced and/or optimal way across RATs and bands based on changes in radio conditions, application requirements, network load, etc. These challenges include maintaining network QoS (e.g., for latency and throughput), configuring carrier aggregation (CA) and/or dual connectivity (DC), bearer type selection (e.g., MCG, SCG, or split bearer for DC), bearer type change for load balancing, etc.

O-RAN traffic steering is specified to be UE-centric rather than cell-centric, based on prediction of UE and network performance. O-RAN traffic steering must be able to select the right set of UEs for a steering action. Thus, the main goal of this use case is to facilitate flexible configuration of optimization policies and performance criteria, which are then used with AI/ML to enable intelligent and proactive traffic management.

SMO provides policy-based and enrichment-based support of ORAN traffic steering. For policy-based support, SMO retrieves necessary performance, configuration, and other data for defining and updating policies to guide behavior of the traffic management function in Near-RT RIC. For example, the policy could specify different optimization objectives to guide the carrier/band preferences at per-UE or group of UE granularity.

For enrichment-based support, SMO retrieves necessary performance, configuration, and other data for performing data statistical analysis to produce enrichment information for Near- RT RIC to assist in the traffic steering function. For example, this could be an analysis method to construct radio fingerprint based on UE measurement report with RSRP/RSRQ/CQI information for serving and neighbor cells.

Advanced analytics systems, such as Ericsson Expert Analytics (EEA), are based on collecting and correlating elementary network events from different network domains, such as core, radio, and transport networks. Such analytics systems calculate user- and session-level E2E service quality metrics (S-KPIs) as well as radio and network resource metrics (R-KPIs) that characterize the radio environment or network operation at user and session level. These types of solutions are suitable for session-based troubleshooting and analysis of network issues.

Event-based analytics require real-time collection and correlation of node and protocol events from different RAN and CN nodes, probing signaling interfaces, and sampling of userplane traffic. Additionally, event-based analytics require an advanced database, a rule engine, and a “big data” analytics platform.

It is expected that 5G networks will serve a higher number of UEs and support a wider variety of service types than 4G and previous-generation networks. This will significantly increase the incoming event rate and type to be processed by analytics systems to support QoS and QoE for users. As such, there is a need for an intelligent approach to focus on events and flows that really need to be collected and correlated.

As mentioned above, when a user initiates a specific service, 5GC assigns that service to a QoS flow (between UE and 5GC, via RAN) that has QoS characteristics (and associated 5QI) that corresponds to the service’s QoS requirements. QoS policy enforcement is performed in the RAN, but without any information about end-user QoE or about classification of the specific services or applications of the end users. Currently, this information is not available to the RAN.

Even though the RAN has QoS information (e.g., 5QI), QoS cannot be translated to QoE without service classification information. Additionally, E2E QoS and end-user QoE cannot be derived based on purely radio events in the RAN. These metrics are based on core network, transport network, and RAN metrics that have been correlated to user identity (e.g., IMSI/SUPI).

Also, the RAN typically relies on time-aggregated metrics such as PM counters for QoS policy enforcement. Due to the averaging or smoothing inherent in these metrics, they do not convey the dynamic fluctuations of data traffic, radio conditions, etc. over time. To summarize, conventional traffic steering solutions are based on cell-level key R-KPIs and/or QoS metrics, and are generally independent of any specific application or service being used by a UE. RAN decisions about steering or switching UE traffic to another cell, RAT, and/or RAN node is independent of specific UE services, which are unknown to the RAN beyond 5 QI provided by 5GC. Thus, there is a need for solutions that better adhere to the O-RAN principles of end-user QoE optimization and user-centric traffic steering.

Embodiments of the present disclosure address these and other problems, issues, and/or difficulties by providing flexible and efficient methods and systems for session-based traffic steering that support end-user QoE optimization. Embodiments are summarized as follows.

Embodiments include an analytics subsystem that correlates R-KPIs, QoS settings, and QoE metrics with network conditions (e.g., load levels) and with service/traffic classification in the network. The correlated information is stored and used to train a system (e.g., AI/ML models) that learns how resources and QoS settings impact end-user QoE, taking into account user mobility, UE type/characteristics, and network slice attributes in addition to service classification.

Embodiments include techniques for correlation of packet core events (PC EBM) and RAN cell trace (CTR) events that provides near-RT service classification binding of traffic/service classification to radio sessions. This can facilitate application of service-specific policies in the RAN.

Embodiments include a learning system (non-RT) that is trained with the correlated information to create service-specific relationships between resources and QoS settings (under control of policies) and achieved QoE metrics. The learning system can create policies based on mapping of the required QoE for a specific service to the necessary resources and QoS settings in the network Service classification and QoS-QoE mapping can be achieved based on CN (e.g., 5GC) and RAN inputs, while O-RAN does not specify use of CN inputs (only more generic "external" inputs).

Embodiments include a session level monitoring function (near-RT), with traffic and service classification capabilities, that applies policies created by the learning system to monitor that conditions necessary for the required end-user QoE are being met. In addition to service classification, this function may consider user- or subscription-based prioritization when selecting the appropriate QoE requirement and policy to be applied in the RAN and CN.

Embodiments include a QoE-preserving traffic steering function that moves UEs across overlapping cells and RATs to meet end-user QoE targets, in view of available network resources. This is performed within a traffic management system based on the service-specific QoS-to-QoE mapping provided by the learning system. Compared to conventional QoS policy enforcement in RANs, embodiments provide at least the following novel and inventive features:

• policies and enforcement are based on E2E QoE instead of QoS targets;

• per-session correlated analytics system is used to collect and correlate service specific resource, QoS, and QoE metric;

• this correlated information is used to train AI/ML models that capture QoS-to-QoE mapping with respect to service classification and network conditions or, in other words, learns the QoS parameters required to meet QoE requirements;

• policies are applied on a per-session basis using session-based monitoring and a sessionbased analytics system, in consideration of service classification, subscription-based prioritization, and optionally network slices and/or UE mobility;

• policies are enforced in RAN and/or CN, depending on what resource conditions are learned to be necessary for meeting QoE requirements;

• traffic steering is based on optimization of derived QoE targets instead of QoS, using the learned QoS-to-QoE mapping, e.g., predicted QoE upon QoS change due to change of RAT, RAN node, etc.; and

• provides the possibility to learn not only impact of observed QoS on QoE (i.e., the QoS- to-QoE mapping), but also impact that applying policies onto the observed QoS has on QoE.

Embodiments can provide various benefits, advantages, and/or solutions to problems described herein. A general goal of any network policy enforcement is maintaining or optimizing user experience or service quality based on controllable network resources. Embodiments provide a currently missing connection between controllable system resources and end-user QoE: an AI/ML model that can predict achieved QoE based on the dedicated resources and QoS settings, with respect to service classification and network conditions. Put differently, policies are based on user- or service-level QoE requirements rather than network QoS metrics.

Additionally, policy decisions are based on a service-specific, per-flow monitoring and analytics system instead of a static QoS framework (e.g., 5QI) as in conventional systems. As such, embodiments regulate QoE with better granularity without requiring overprovisioning of network resources to provide safety margins, as is done in conventional QoS-based techniques. Thus, embodiments can facilitate more efficient use of network resources and reduce network load. Moreover, embodiments can consider additional information for policy enforcement than conventional techniques, such as service classification, subscriber data, user mobility, UE type, etc. Additionally, embodiments provide user-centric QoE optimization and traffic steering in contrast to conventional cell-centric solutions that are based on network resources and QoS without regard to actual user experience. In this manner, embodiments are better at enforcing service level agreements (SLAs) that have the ultimate goal of end-user experience.

Two fundamental elements of embodiments of the present disclosure are service-specific QoE enforcement and IMSI- and service-level resolution of policies. Note that IMSI is an abbreviation for international mobile subscriber identifier (IMSI), which uniquely identifies each user based on network subscription.

In general, service-specific QoE enforcement requires a “missing link" between radio resource control and application-level QoE metrics. This can be provided by the Non-RT RIC in combination with a novel session-level E2E correlation capability (e.g., within or outside of the SMO). An AI/ML model is trained based on collected samples of past traffic. During training of the model, the policy control options (e.g., radio resources, QoS parameters) and the service/traffic classification are the input “features” and the corresponding achieved QoE is the output “label”. The model can be retrained continuously, periodically, or occasionally to learn how parameters controlled by the policies impact QoE, and can generate a policy for resource and QoS settings that produce the desired QoE.

The achieved QoE depends on the specific service being used; hence, service and traffic classification is important. E2E correlation creates a mapping between network conditions and the observed QoE on a per-session basis. As services and their sensitivity to network conditions changes over time, the AI/ML model in Non-RT RIC can be retrained and updated accordingly, with corresponding policy updates being provided to Near-RT RIC over the Al interface. The Near-RT RIC can manage RAN and resources to maintain QoE according to these policies.

Due to the service-specific QoE-to-QoS/resource mapping (instead of conventional service-independent QoS control), the policies need service and traffic classification as inputs for the decisions (e.g., for model inference). This information is provided by associated IMSIs (or other unique UE or subscription identifier) with per-flow monitoring actions. IMSI association also provides the capability for subscription specific prioritization in addition to service-specific prioritization and/or management.

In general, embodiments include functions for QoE optimized policy enforcement and QoE-preserving traffic steering, which align with O-RAN use cases 4-5 discussed above. Figure 4 shows a high-level block diagram of a system (400) for session-based traffic steering with end-user QoE optimization, according to embodiments of the present disclosure. Note that the functions for QoE optimized policy enforcement and QoE-preserving traffic steering are shown as blocks that interface to a “pipeline” of enabler blocks arranged horizontally in the middle of Figure 4. At a high level, this pipeline collects RAN and CN (e.g., 5GC) data on user data sessions, performs session-level service classification, and correlates RAN/CN events for each session classified in this manner.

The correlated and classified session data is then used for various purposes. For example, the session data is used to train an AI/ML model, with acquired QoE metrics as labels and correlated QoS metrics (and R-KPIs) as features. The resulting AI/ML model can then be used for inference of QoE from available QoS/R-KPI values. The AI/ML models can be translated to service-specific policies for QoS requirements according to desired QoE.

These can be provided to a policy controller (e.g., a library or database) and used for QoE optimized policy enforcement. In this manner, policy enforcement relies not only on R- KPIs and QoS metrics but also on service classification. Service-specific policies are applied to achieve service-specific QoE targets, and can be enforced not only in the RAN but also in the CN, as needed.

The trained AI/ML models are also used for traffic steering, e.g., when moving sessions across various cells, RATs, and RAN nodes in accordance with service-specific QoE requirements and in view of network resource efficiency. Since QoE depends on the service itself, service classification and session-level correlated data is also sued for QoE-preserving traffic steering.

Figure 5 is a more detailed block diagram that illustrates an exemplary configuration of the blocks/functions shown in Figure 4 in the O-RAN architecture. In Figure 5, these blocks/ functions are labelled collectively as QoE-based network optimizer (500) and are surrounded by a solid line, whereas dashed-line boxes are used to represent O-RAN architectural elements.

As Figure 5 shows, the ML Training Module and Policy Controller functions reside in the SMO while Policy Enforcement and Traffic Steering functions reside in the Non-RT RIC and receive trained AI/ML models over the R1 interface. The Correlation and Service Classification functions are shown as residing in the SMO but can also reside outside the SMO as an external system.

The data collection function is not shown but can reside together with Correlation and Service Classification, since those functions use the collected data. In particular, the system collects cell trace (CTR) and PM counter data from the RAN (510, e.g., NG-RAN) as well as User Plane (UP) and Control Plane (CP) events from CN (520, e.g., 5GC), and optionally data from other service-related data sources (e.g., IP Multimedia Sub-system, IMS). The RAN (510) and CN (520) comprise a communication network (530, e.g., a 5G network).

Figure 6 shows a more detailed diagram of a system (600) configured for session-based traffic steering with end-user QoE optimization, according to embodiments of the present disclosure. In particular, Figure 6 shows various sub-blocks or sub-functions and data flow among these, which are described below.

A UE-level (or IMSI-level) analytics block includes the Data Correlation (or Correlator), Service Classifier, and Flow-level QoE Calculation blocks. The Correlator collects all RAN data and UP data associated with each user data session into a single record, both to learn how observed QoS translates to QoE for specific services and to create a binding between identifiers in the UP (where service classification happens) and in the RAN (where policies are enforced). The correlated records can also include other parameters such as IMSI, international mobile equipment identifier (IMEI), 5QI, RAT, UE location information (e.g., cell ID and/or coordinates), etc. The Correlator maps the collected RAN and UP data to IMSI (per session) in near-RT (e.g., every minute as shown).

The IMSI-level analytics block also performs service classification based on the collected UP data and calculates RAN and UP KPIs based on the correlated session data in each record. UP KPIs include QoS parameters such delay, jitter, packet loss, etc. RAN KPIs include serving and neighbor cell radio measurements (e.g., RSRP, RSRQ), cell load, radio-related events (e.g., handover success/failure, handover time, etc.). For example, AI/ML models can be used to predict flow level QoS KPI metrics at the last stage of the IMSI-level analytics workflow. QoS metrics can be collected from the service-specific data sources (e.g., IMS) and/or calculated from available data such as video or voice quality or similar service experience metrics.

Subsequently, the correlated data and the QoS metrics are sent to the Learning System block. At this point, information from additional sources (i.e., other than RAN and CN) can be correlated to IMSIs, such as a user profile, UE capabilities (e.g., type, model, OS, etc.), or other information that is not collected directly from RAN/CN but associated with user or UE identifiers such as IMSI or IMEI.

The QoE Calculation function forecasts or predicts per-UE QoE (e.g., QoE KPI) based on the observed QoS metrics considering all the information from the correlated user session data and additional cell performance information from PM counters. This can be done using any regression-type AI/ML model that restricts model outputs to their maximum/minimum of image set where they are represented. An exemplary model would be a gradient boosting or restricted linear regression for Video MOS (QoE), which is estimated from low level QoS metrics such as bitrate, video stall time, time to video initialization, video resolution changes during the session, time spent on different video resolutions, etc.

For example, a regression-type AI/ML model uses weighted average (or other statistics) of service-specific QoE metrics as target variables for regression. It predicts or forecasts servicespecific QoE (e.g., audio MOS) based on UE-specific correlated data from flow-level QoS metrics, e.g., using gradient boosted trees. For translation between QoE and QoS, this process needs data with defined (e.g., labelled) relationships between QoE and QoS, such as previously determined by the system for other users or obtained based on predefined rules, etc.

As a more specific example, the following linear regression can be used:

where:

• Y is a vector of different QoE value measurements;

• X is a matrix of different QoE KPI measurements, with each column i containing different KPIs and each row j containing a new instance of measurements;

• B is a matrix of different weights of QoE KPI measurements, where Bij denotes the weight of the j -th measurement of the i-th KPI; and

• e is the squared error/residual of the regression on j-th measurement.

In addition to mapping QoS metrics to QoE, the system also must perform the inverse mapping of QoE to QoS metrics to support selecting the most appropriate policy to for a desired user- and service-specific QoE. In other words, the system must identify QoS constraints or requirements from QoE, then find the available policy that best supports those QoS constraints. A generic approach that is agnostic to type and structure of AI/ML model is binary search, moving in the space of QoS vectors and checking the corresponding QoE to identify QoS vectors that provide the desired QoE. More sophisticated approaches can utilize the internal structure of the trained AI/ML model (e.g., internal splits of gradient boosted trees) to distinguish between QoS vectors that support different QoE regressions and identify QoS vectors that provide the desired QoE. Another possible approach is using an AI/ML model explainer, which can select some input data (e.g., QoS data) that is explanatory of output results (e.g., QoE).

The Learning System also includes an ML Lifecycle Management (LCM) function that observes the quality of the current QoE calculation based on the AI/ML model to detect model drift during dynamically changing conditions. If sufficient model drift is detected, the ML LCM function retrains the model and distributes the retrained model for use by other parts of the system, as discussed below.

For example, model drift can be detected based on observing model outputs and inputs, e.g., by binning the data by discrete intervals (also called “categories”) using a histogram and observing the input data drift by the drift over time in the distribution per category. Distribution among categories can be compared for different time periods (e.g., last hour, last day, etc.) to see if is noticeably changing, which indicates drift and possibly need for retraining. As an implementation example, drift detection can be performed by a fixed cumulative window method (FCWM) that compares distribution of a current window to the fixed determined window (e.g., window of training data) using Kullback-Leibler hypothesis testing.

In some embodiments, the Learning System also includes a Reinforcement Learning of Policy function that identifies and learns both the direct impact of observed QoS conditions on achieved QoE as well indirect impact of QoE-based policies on achieved QoE based on the observed QoS. For example, if there are policies for maximum latency, minimum throughput, etc., users will not always experience those exact values. Reinforcement Learning can be used to learn actual QoE experienced by users based on policies associated with QoE minimum/maximum boundaries such as these.

For example, Reinforcement Learning of Policy can be implemented as a Markov Decision Process (MDP) module that runs in the backpropagation feedback loop, learning the policies iteratively and self-tuning them dynamically with the changes of the network (or data). As a more specific example, this module can use the backpropagated policies and their effect on the network, and creates a greedy algorithm that applies the best known policy (“target policy”) on almost all RAN nodes and another learned policy (“behavior policy”) on a small subset of RAN nodes. Put mathematically, the algorithm selects with probability e the best known (or greedy) policies 7T « 7T* and with probability 1 — 6 a random other action or policy. This learning behavior is also known to be 6-soft. In this example, the Learning system learns on additional policy configuration data collected, in a Monte Carlo-type approach, by searching the available policy space for a more optimal policy while minimizing (or tightly controlling) the impact on actual QoE delivered by the network.

Reinforcement Learning of Policy can also be viewed as learning the best available policy- driven action (e.g., prioritizing a user session, passing the user session to another cell, making no change to the user session) for each network state (e.g., distribution of users and QoE requirements). For example, actions can be given a positive score when they increase average QoE within a cell and a negative score when they decrease average QoE within a cell. This can be used to identify both locally optimal policy actions as well as globally optimal policy actions.

In this manner, traffic steering actions (discussed more below) and resulting servicespecific QoE measurements can be learned and used to modify subsequent traffic steering actions. For example, congestion mitigation is done by moving many UEs using texting/chat services from one cell to another. Resulting QoE measurements of throughput indicate that congestion was not significantly reduced, and QoE measurements of reliability indicate that these texting/chat services went down due to a high number of handovers during congestion mitigation. In such case, the Learning System learns that texting/chat services should not be part of load balancing handovers to mitigate “throughput degraded” congestion. However, the Learning System could produce a policy for “too many users” congestion mitigation that involves load balancing handovers of UEs using texting/chat services.

The Learning System outputs service-specific policies to the Policy and Charging Rules Function (PCRF), which is responsible for policy enforcement in the network. Additionally, the Learning System outputs service-specific QoS-to-QoE mappings to the PCRF, which can be viewed as a “policy library”.

Conventionally, PCRF selects policies based on 5QI (or similar QoS-related information) and perhaps user subscription type. In embodiments of the present disclosure, PCRF also selects policies based on the service being used, which was previously identified by the service classification function in UE-level analytics. In various embodiments, servicespecific, QoE-optimizing policy enforcement can be a default mode of operation or can be a non-default, event-triggered mode of operation. For example, PCRF can initiate this non-default mode when PM counters indicated negative trends in a cell, cell resource utilization exceeds a threshold, etc.

Once service-specific, QoE-optimizing policy enforcement is applied, the PCRF obtains UE-specific service information and selects a policy to facilitate meeting the service- specific QoE requirements. For example, the selected policy can be a set of QoS requirements that results in the RAN/CN meeting the UE- and service-specific QoE requirements. This QoS-to-QoE mapping was established during AI/ML model training. In some embodiments, after the policy is selected, it is backpropagated to the Reinforcement Learning function discussed above.

QoE-preserving traffic steering is based on the service classifications and the QoS-to- QoE mapping provided by other blocks or functions in Figure 6. This traffic steering is typically triggered by network PM metrics reaching some threshold, indicating overload of cells, RAN nodes, etc. Priority to switching one or more UEs, the traffic steering first determines expected QoS for the UEs in a candidate cell, node, RAT, etc. Compared to conventional techniques, embodiments translate the expected QoS to QoE for the specific service used by the UEs, which is known from the service classification. Alternately, this process can be viewed as using an “inversion” of the trained AI/ML model(s), to identify QoS settings (output) needed to meet QoE requirements (input) for the specific service used by the UEs, then determining if the candidate(s) satisfy those QoS settings.

One example application for embodiments of the present disclosure is different policies for mobile and static users. Conventionally, when a large (“umbrella’) 4G cell becomes congested, all traffic associated with a 5QI class may be redirected to a high- or mid-band 5G cell or a small capacity 4G cell to resolve the congestion. However, this policy may be suboptimal for mobile UEs since it generates more handovers, which can cause additional QoE degradation and generate excess signaling load.

In embodiments of the present disclosure, the per-flow location history indicates whether the UE is static or mobile. As such, policy to direct traffic to a high- or mid-band 5G cell or a small capacity 4G cell is applied only to static UEs only, while mobile UEs are served by the low- band 4G umbrella cell, thereby avoiding intra-RAT and intra-frequency handovers of those UEs.

Another example application for embodiments of the present disclosure is different policies for different subscription types or profiles. In general, there are different subscription profiles in each PLMN, including “VIP” subscribers who should receive better and/or higher- priority service (e.g., higher QoE) than other, non- VIP subscribers. Service quality degradation is experienced in a cell due to high number of active subscribers (cell congestion), which affects VIP and non- VIP subscribers, even though VIP subscriber traffic for certain services may receive higher priority.

Conventionally, when a cell becomes congested, a random portion of UEs connected to that cell are forced to handover to another (e.g., overlapping or neighbor) cell to relieve the congested cell. However, there is no guarantee that the other cell provides an appropriate service quality or that the handover to the other cell may fail and ruin the QoE of the VIP subscriber.

Based on the IMSI-aware per-flow correlated records, embodiments of the analytics system are able to steer traffic separately for VIP and non- VIP subscribers based on QoE predictions. For example, an exemplary policy could be that non- VIP subscribers are handed over to overlapping cells while VIP subscribers are kept in the original cell. By relieving the congested cell in this manner, the VIP subscribers receive the appropriate service quality and do not risk handover failure. Normal subscribers can receive a better but less guarantied service quality in the overlapping cells, but at some risk of handover failure.

In conventional solutions, policies are applied for a QoS flow only when QoS parameter degradation is experienced. Assume that packet loss for a 5QI class exceeds a threshold that triggers a policy change, including changing the QoS profile of the cell. As a consequence, more radio resources (e.g., physical resources blocks, PRBs) are allocated to the affected 5QI traffic class, which reduces the packet loss but has the negative effect of decreasing cell capacity.

QoS degradation can affect user QoE for different services in different ways. For example, increased packet loss can cause QoE degradation for over-the-top (OTT) voice applications such as Teams, Messenger, etc. but may cause no QoE degradation for other services such as web browsing, messaging, file transfer, etc. - even if those services are associated with the same 5QI class. As such, an observed QoS degradation in a cell does not necessarily produce QoE degradation for specific services even if those services are used by many subscribers in the cell. If QoE for these specific services are not affected by the observed QoS degradation, or there no QoE-sensitive services being used in the cell, then it is unnecessary to allocate more radio resources based on the QoS degradation.

In embodiments of the present disclosure, the per- flow records can include application ID and application type for each session. Based on this information, a policy to allocate an affected session to a different 5QI class can be restricted to only services with QoE metrics sensitive to the particular QoS degradation, which can improve or maintain QoE for these services while maintaining radio resource efficiency.

As currently specified, the O-RAN architecture does not include any components and/or interfaces that enable input data flows from existing data collection components for crossdomain correlation. For example, the SMO Non-RT RIC component does not have any data interfaces towards domains other than RAN. More generally, input data from non-RAN domains (e.g., CN, Application, etc.) are out of scope of O-RAN. Even so, two possible implementation options for integrating embodiments of the present disclosure into O-RAN architecture are described below.

Figure 7 shows a first implementation option for integrating a system (700) according to embodiments of the present disclosure into O-RAN architecture. In this option, the per-flow analytics cross-domain correlator component (710) runs on an Al server outside of SMO (e.g., on public or private cloud computing environment), and has an external interface into SMO. The per-flow analytics cross-domain correlator component also has external interfaces that facilitate data collection from other domains such as CN (e.g., 5GC), IMS, etc. The Non-RT RIC (720) in the SMO can configure network resources based on policies and/or models provided by the per-flow analytics component, similar to the arrangement shown in Figure 5.

Figure 8 shows a second implementation option for integrating a system (800) according to embodiments of the present disclosure into O-RAN architecture. In this option, the per-flow analytics cross-domain correlator component (820) runs inside SMO, which has external interfaces that facilitate data collection from other domains such as CN (e.g., 5GC), IMS, etc. and from other data collecting devices. Similar to Figure 7, the Non-RT RIC (820) in the SMO can configure network resources based on policies and/or models provided by the per-flow analytics component.

These two options are harmonized with various O-RAN proposals, such as the statement that “Non-RT RIC will invoke the corresponding training model/application in an Al server inside SMO (it can be placed outside SMO also)” in relation to use case 16.

Various features of the embodiments described above correspond to various operations illustrated in Figure 9 (including parts A and B), which shows an exemplary method e.g., procedure) for managing communication network resources based on end-user QoE, according to various embodiments of the present disclosure. In other words, various features of the operations described below correspond to various embodiments described above. Although Figure 9 shows specific blocks in a particular order, the operations of the exemplary method can be performed in a different order than shown and can be combined and/or divided into blocks having different functionality than shown. Optional blocks or operations are indicated by dashed lines.

The following description is based on the exemplary method being performed by a network management system for the communication network. As a more specific example, the network management system can be an SMO system. However, the exemplary method can be performed by other nodes, functions, or systems within or outside of the communication network. Furthermore, one or more operations of the exemplary method can be performed by a particular function of an SMO system, such as a non-RT RIC.

The exemplary method can include the operations of block 910, where the network management system can collect performance -related data from a RAN (e.g., an O-RAN) and a core network (CN) comprising the communication network. The exemplary method can also include the operations of block 930, where the network management system can correlate at least the collected performance-related data into a plurality of records corresponding to a plurality of user data sessions with the communication network. The exemplary method can also include the operations of block 940, where the network management system can train one or more AI/MF models based on the correlated data records, wherein each AI/MF model maps between network QoS metrics and service-specific QoE metrics. The exemplary method can also include the operations of block 960, where the network management system can configure communication network resources to carry one or more second user data sessions, based on the trained AI/ML model and on QoE requirements of a service comprising the second user data sessions.

In some embodiments, the performance-related data (e.g., collected in block 910) includes one or more of the following:

• trace data for respective cells provided by RAN nodes;

• performance management (PM) counter data associated with the RAN nodes;

• user plane (UP) event information associated with the CN; and

• control plane (CP) event information associated with the CN.

In some of these embodiments, the UP event information includes information that identifies one or more of the following:

• respective user data sessions;

• user subscriptions associated with the respective user data sessions;

• UEs associated with the respective user data sessions; • services associated with the respective user data sessions; and

• network QoS metrics associated with at least one of the following: the respective user data sessions, and the services associated with the respective user data sessions.

In some embodiments, each correlated data record includes one or more of the following associated with the corresponding user data session:

• an identifier of a service;

• an identifier of user network subscription;

• identifier of a UE;

• location of the UE;

• mobility history of the UE;

• radio access technology (RAT) being used in the RAN ; and

• an identifier corresponding to a plurality of network QoS characteristics.

In some embodiments, correlating at least the collected performance-related data in block 930 includes the operations of sub-block 931, where the network management system can determine, from the collected performance-related data, the following network QoS metrics associated with each user data session: one or more RAN QoS metrics, and one or more CN QoS metrics. In some of these embodiments, the RAN QoS metrics associated with each user data session include one or more of the following: RAN resources used, serving cell load, mobility events between serving cells, and serving and neighbor cell radio measurements. In some of these embodiments, the CN QoS metrics associated with each user data session include one or more of the following: packet delay, packet delay jitter, packet loss, and priority level.

In some of these embodiments, the plurality of user data sessions are associated with a corresponding plurality of services and the exemplary method can also include the operations of block 920, where the network management system can collect service-specific QoE data for the plurality of user data sessions. In such case, correlating at least the collected performance-related data in block 930 can also include the operations of sub-block 932, where the network management system can correlate the collected service-specific QoE data into the plurality of data records.

In some embodiments, each AI/ML model is trained based on the correlated data records using the network QoS metrics as input features and the service-specific QoE data as output labels. In some variants, the service-specific QoE data includes one or more of the following: opinion score, service availability, bitrate statistics, playback errors, rebuffering, startup time, startup failures, and duration of user data session.

In some embodiments, the exemplary method can also include the operations of block 950, where the network management system can determine a plurality of service-specific policies for network resource management, based on the one or more trained AI/ML models and on QoE requirements of a respective plurality of services. Each service-specific policy includes corresponding network QoS requirements.

In some of these embodiments, configuring communication network resources to carry one or more second user data sessions in block 960 includes the operations of sub-block 961, where the network management system can enforce service-specific policies for the respective second user data sessions. In some variants, enforcing the service-specific policies in sub-block 961 can include the following operations for each second user data session, labelled with corresponding sub-sub-block designations:

• (961a) determining network QoS requirements for the second user data session based on the one or more trained AI/ML models and on QoE requirements of a service comprising the second user data session;

• (961b) selecting one of the service-specific policies that includes network QoS requirements that meet the network QoS requirements for the second user data session; and

• (961c) configuring communication network resources for the second user data session in accordance with the network QoS requirements included in the selected service-specific policy.

In some of these embodiments, the exemplary method also includes the operations of block 970, where the network management system can collect one or more of the following feedback for the respective second user data sessions: observed network QoS metrics associated with the configured communication resources, and observed QoE for the respective services. In some variants, the exemplary method can also include the operations of block 980, where the network management system can retrain the one or more AI/ML models based on the collected feedback.

In other variants of these embodiments, a first service-specific policy is selected for a first subset and a second service-specific policy is selected for a second subset of the second user data sessions (i.e., for which feedback was collected). In such variants, the exemplary method can also include the operations of block 990, where the network management system can perform reinforcement learning (RL) for policy selection based on the collected feedback for the second user data sessions. An example of these variants is the Markov Decision Process (MDP) discussed above.

In some embodiments, configuring communication network resources to carry one or more second user data sessions in block 960 can include the operations of sub-block 965, where the network management system can change one or more network resources or settings for each second user data session based on the respective QoE requirements of the services comprising the second user data sessions. In some of these embodiments, the one or more network resources or settings (i.e., that are changed) include one or more of the following: serving cell, resource allocation within serving cell, serving radio access technology (RAT), serving RAN node, and QoS settings. An example of these embodiments is the “traffic steering” functionality discussed above.

In some of these embodiments, configuring communication network resources to carry one or more second user data sessions in block 960 can include the operations of sub-block 962, where while a plurality of user data sessions are active, the network management system can detect a change in network QoS metrics indicating a degradation in communication network performance. In some variants, configuring communication network resources to carry one or more second user data sessions in block 960 also includes the following operations labelled with corresponding sub-block numbers:

• (963) using the one or more AI/ML models, predicting QoE for the respective services comprising the active user data sessions based on the detected change in network QoS metrics; and

• (966) refraining from changing network resources or settings for one or more third user data sessions comprising services having predicted QoE that meet corresponding QoE requirements.

Furthermore, the one or more second user data sessions (whose network resources or settings are changed) comprise services having predicted QoE that do not meet corresponding QoE requirements.

In some further variants, configuring communication network resources to carry one or more second user data sessions in block 960 can include the operations of sub-block 964, where using the one or more AI/ML models, the network management system can predict further QoE for the respective services comprising the second user data sessions based on a further change in network QoS metrics associated with the changed network resources or settings. In such case, the predicted further QoE meet the corresponding QoE requirements, which can trigger the network management system to perform the change in network resources or settings for the second user data sessions.

In other further variants, the one or more second user data sessions (whose network resources or settings are changed) are associated with one or more of the following: lower-priority user subscription profiles (e.g., “non-VIP” users) and lower- mobility users. In such variants, configuring communication network resources to carry one or more second user data sessions in block 960 can also include the operations of sub-block 967, where the network management system can refrain from changing network resources or settings for one or more third user data sessions associated with one or more of the following: higher-priority user subscription profiles (e.g., “VIP” users) and higher-mobility users.

In some embodiments, configuring communication network resources in block 910 can be performed by a non-real-time RAN intelligent controller of an SMO system, and the collecting, correlating, and training operations in blocks 930, 940, and 960 respectively are performed by the SMO system (i.e., outside of the non-RT RIC). In other embodiments, the collecting, correlating, and training operations in blocks 930, 940, and 960 can be performed by a cloud computing environment external to the SMO system.

Although various embodiments are described above in terms of methods, techniques, and/or procedures, the person of ordinary skill will readily comprehend that such methods, techniques, and/or procedures can be embodied by various combinations of hardware and software in various systems, communication devices, computing devices, control devices, apparatuses, non-transitory computer-readable media, computer program products, etc.

Figure 10 shows an example of a communication system 1000 in accordance with some embodiments. In this example, the communication system 1000 includes a telecommunication network 802 that includes an access network 1004, such as a RAN, and a core network 1006, which includes one or more core network nodes 1008. In some embodiments, telecommunication network 802 can also include one or more Network Management (NM) nodes 1018, which can be part of an operation support system (OSS) or a business support system (BSS). The NM nodes can monitor and/or control operations of other nodes in access network 1004 and core network 1006. Although not shown in Figure 10, NM node 1018 is configured to communicate with other nodes in access network 1004 and core network 1006 for these purposes.

Access network 1004 includes one or more access network nodes, such as network nodes 1010a and 1010b (one or more of which may be generally referred to as network nodes 1010), or any other similar 3GPP access node or non-3GPP access point. The network nodes 1010 facilitate direct or indirect connection of UEs, such as by connecting UEs 1012a, 1012b, 1012c, and 1012d (one or more of which may be generally referred to as UEs 1012) to the core network 1006 over one or more wireless connections.

Example wireless communications over a wireless connection include transmitting and/or receiving wireless signals using electromagnetic waves, radio waves, infrared waves, and/or other types of signals suitable for conveying information without the use of wires, cables, or other material conductors. Moreover, in different embodiments, the communication system 1000 may include any number of wired or wireless networks, network nodes, UEs, and/or any other components or systems that may facilitate or participate in the communication of data and/or signals whether via wired or wireless connections. The communication system 1000 may include and/or interface with any type of communication, telecommunication, data, cellular, radio network, and/or other similar type of system.

The UEs 1012 may be any of a wide variety of communication devices, including wireless devices arranged, configured, and/or operable to communicate wirelessly with the network nodes 1010 and other communication devices. Similarly, the network nodes 1010 are arranged, capable, configured, and/or operable to communicate directly or indirectly with the UEs 1012 and/or with other network nodes or equipment in the telecommunication network 1002 to enable and/or provide network access, such as wireless network access, and/or to perform other functions, such as administration in the telecommunication network 1002.

In the depicted example, the core network 1006 connects the network nodes 1010 to one or more hosts, such as host 1016. These connections may be direct or indirect via one or more intermediary networks or devices. In other examples, network nodes may be directly coupled to hosts. The core network 1006 includes one more core network nodes (e.g., core network node 1008) that are structured with hardware and software components. Features of these components may be substantially similar to those described with respect to the UEs, network nodes, and/or hosts, such that the descriptions thereof are generally applicable to the corresponding components of the core network node 1008. Example core network nodes include functions of one or more of a Mobile Switching Center (MSC), Mobility Management Entity (MME), Home Subscriber Server (HSS), Access and Mobility Management Function (AMF), Session Management Function (SMF), Authentication Server Function (AUSF), Subscription Identifier De-concealing function (SIDE), Unified Data Management (UDM), Security Edge Protection Proxy (SEPP), Network Exposure Function (NEF), and/or a User Plane Function (UPF).

Host 1016 may be under the ownership or control of a service provider other than an operator or provider of the access network 1004 and/or the telecommunication network 1002, and may be operated by the service provider or on behalf of the service provider. The host 1016 may host a variety of applications to provide one or more service. Examples of such applications include live and pre-recorded audio/video content, data collection services such as retrieving and compiling data on various ambient conditions detected by a plurality of UEs, analytics functionality, social media, functions for controlling or otherwise interacting with remote devices, functions for an alarm and surveillance center, or any other such function performed by a server. As a specific example, host 1016 can be implemented in a cloud computing environment.

In some embodiments, access network 1004 can include a service management and orchestration (SMO) system or node 1020, which can monitor and/or control operations of the access network nodes 1010. This arrangement can be used, for example, when access network 1004 utilizes an Open RAN (O-RAN) architecture. SMO system 1020 can be configured to communicate with core network 1006 and/or host 1016, as shown in Figure 10.

In some embodiments, one or more of host 1016, network management node 1018, and SMO system 1020 can be configured to perform various operations of exemplary methods (e.g., procedures) for managing communication network resources based on end-user QoE, such as describe above in relation to Figure 9.

As a whole, the communication system 1000 of Figure 10 enables connectivity between the UEs, network nodes, and hosts. In that sense, the communication system may be configured to operate according to predefined rules or procedures, such as specific standards that include, but are not limited to: Global System for Mobile Communications (GSM); Universal Mobile Telecommunications System (UMTS); Long Term Evolution (LTE), and/or other suitable 2G, 3G, 4G, 5G standards, or any applicable future generation standard (e.g., 6G); wireless local area network (WLAN) standards, such as the Institute of Electrical and Electronics Engineers (IEEE) 802.11 standards (WiFi); and/or any other appropriate wireless communication standard, such as the Worldwide Interoperability for Microwave Access (WiMax), Bluetooth, Z-Wave, Near Field Communication (NFC) ZigBee, LiFi, and/or any low-power wide-area network (LPWAN) standards such as LoRa and Sigfox.

In some examples, the telecommunication network 1002 is a cellular network that implements 3GPP standardized features. Accordingly, the telecommunications network 1002 may support network slicing to provide different logical networks to different devices that are connected to the telecommunication network 1002. For example, the telecommunications network 1002 may provide Ultra Reliable Low Latency Communication (URLLC) services to some UEs, while providing Enhanced Mobile Broadband (eMBB) services to other UEs, and/or Massive Machine Type Communication (mMTC)/Massive loT services to yet further UEs.

In some examples, the UEs 1012 are configured to transmit and/or receive information without direct human interaction. For instance, a UE may be designed to transmit information to the access network 1004 on a predetermined schedule, when triggered by an internal or external event, or in response to requests from the access network 1004. Additionally, a UE may be configured for operating in single- or multi-RAT or multi-standard mode. For example, a UE may operate with any one or combination of Wi-Fi, NR (New Radio) and LTE, i.e., being configured for multi-radio dual connectivity (MR-DC), such as E-UTRAN (Evolved-UMTS Terrestrial Radio Access Network) New Radio - Dual Connectivity (EN-DC).

In the example, the hub 1014 communicates with the access network 1004 to facilitate indirect communication between one or more UEs (e.g., UE 1012c and/or 1012d) and network nodes (e.g., network node 1010b). In some examples, the hub 1014 may be a controller, router, content source and analytics, or any of the other communication devices described herein regarding UEs. For example, the hub 1014 may be a broadband router enabling access to the core network 1006 for the UEs. As another example, the hub 1014 may be a controller that sends commands or instructions to one or more actuators in the UEs. Commands or instructions may be received from the UEs, network nodes 1010, or by executable code, script, process, or other instructions in the hub 1014. As another example, the hub 1014 may be a data collector that acts as temporary storage for UE data and, in some embodiments, may perform analysis or other processing of the data. As another example, the hub 1014 may be a content source. For example, for a UE that is a VR headset, display, loudspeaker or other media delivery device, the hub 1014 may retrieve VR assets, video, audio, or other media or data related to sensory information via a network node, which the hub 1014 then provides to the UE either directly, after performing local processing, and/or after adding additional local content. In still another example, the hub 1014 acts as a proxy server or orchestrator for the UEs, in particular in if one or more of the UEs are low energy loT devices.

The hub 1014 may have a constant/persistent or intermittent connection to the network node 1010b. The hub 1014 may also allow for a different communication scheme and/or schedule between the hub 1014 and UEs (e.g., UE 1012c and/or 1012d), and between the hub 1014 and the core network 1006. In other examples, the hub 1014 is connected to the core network 1006 and/or one or more UEs via a wired connection. Moreover, the hub 1014 may be configured to connect to an M2M service provider over the access network 1004 and/or to another UE over a direct connection. In some scenarios, UEs may establish a wireless connection with the network nodes 1010 while still connected via the hub 1014 via a wired or wireless connection. In some embodiments, the hub 1014 may be a dedicated hub - that is, a hub whose primary function is to route communications to/from the UEs from/to the network node 1010b. In other embodiments, the hub 1014 may be a non-dedicated hub - that is, a device which is capable of operating to route communications between the UEs and network node 1010b, but which is additionally capable of operating as a communication start and/or end point for certain data channels.

Figure 11 shows a network node 1100 in accordance with some embodiments. As used herein, network node refers to equipment capable, configured, arranged and/or operable to communicate directly or indirectly with a UE and/or with other network nodes or equipment, in a telecommunication network. Examples of network nodes include, but are not limited to, access points (APs) (e.g., radio access points), base stations (BSs) (e.g., radio base stations, Node Bs, evolved Node Bs (eNBs) and NR NodeBs (gNBs)).

Base stations may be categorized based on the amount of coverage they provide (or, stated differently, their transmit power level) and so, depending on the provided amount of coverage, may be referred to as femto base stations, pico base stations, micro base stations, or macro base stations. A base station may be a relay node or a relay donor node controlling a relay. A network node may also include one or more (or all) parts of a distributed radio base station such as centralized digital units and/or remote radio units (RRUs), sometimes referred to as Remote Radio Heads (RRHs). Such remote radio units may or may not be integrated with an antenna as an antenna integrated radio. Parts of a distributed radio base station may also be referred to as nodes in a distributed antenna system (DAS).

Other examples of network nodes include multiple transmission point (multi-TRP) 5G access nodes, multi-standard radio (MSR) equipment such as MSR BSs, network controllers such as radio network controllers (RNCs) or base station controllers (BSCs), base transceiver stations (BTSs), transmission points, transmission nodes, multi-cell/multicast coordination entities (MCEs), Operation and Maintenance (O&M) nodes, Operations Support System (OSS) nodes, Self-Organizing Network (SON) nodes, network management nodes, service management and orchestration (SMO) nodes, positioning nodes (e.g., Evolved Serving Mobile Location Centers (E-SMLCs)), and/or Minimization of Drive Tests (MDTs).

In some embodiments, network node 1100 can be configured to perform various operations of exemplary methods e.g., procedures) for managing communication network resources based on end-user QoE, such as describe above in relation to Figure 9.

The network node 1100 includes a processing circuitry 1102, a memory 1104, a communication interface 1106, and a power source 1108. The network node 1100 may be composed of multiple physically separate components (e.g., a NodeB component and a RNC component, or a BTS component and a BSC component, etc.), which may each have their own respective components. In certain scenarios in which the network node 1100 comprises multiple separate components (e.g., BTS and BSC components), one or more of the separate components may be shared among several network nodes. For example, a single RNC may control multiple NodeB s. In such a scenario, each unique NodeB and RNC pair, may in some instances be considered a single separate network node. In some embodiments, the network node 1100 may be configured to support multiple radio access technologies (RATs). In such embodiments, some components may be duplicated (e.g., separate memory 1104 for different RATs) and some components may be reused (e.g., a same antenna 1110 may be shared by different RATs). The network node 1100 may also include multiple sets of the various illustrated components for different wireless technologies integrated into network node 1100, for example GSM, WCDMA, LTE, NR, WiFi, Zigbee, Z-wave, LoRaWAN, Radio Frequency Identification (RFID) or Bluetooth wireless technologies. These wireless technologies may be integrated into the same or different chip or set of chips and other components within network node 1100. The processing circuitry 1102 may comprise a combination of one or more of a microprocessor, controller, microcontroller, central processing unit, digital signal processor, application-specific integrated circuit, field programmable gate array, or any other suitable computing device, resource, or combination of hardware, software and/or encoded logic operable to provide, either alone or in conjunction with other network node 1100 components, such as the memory 1104, to provide network node 1100 functionality.

In some embodiments, the processing circuitry 1102 includes a system on a chip (SOC). In some embodiments, the processing circuitry 1102 includes one or more of radio frequency (RF) transceiver circuitry 1112 and baseband processing circuitry 1114. In some embodiments, the radio frequency (RF) transceiver circuitry 1112 and the baseband processing circuitry 1114 may be on separate chips (or sets of chips), boards, or units, such as radio units and digital units. In alternative embodiments, part or all of RF transceiver circuitry 1112 and baseband processing circuitry 1114 may be on the same chip or set of chips, boards, or units.

The memory 1104 may comprise any form of volatile or non-volatile computer-readable memory including, without limitation, persistent storage, solid-state memory, remotely mounted memory, magnetic media, optical media, random access memory (RAM), read-only memory (ROM), mass storage media (for example, a hard disk), removable storage media (for example, a flash drive, a Compact Disk (CD) or a Digital Video Disk (DVD)), and/or any other volatile or non-volatile, non-transitory device-readable and/or computer-executable memory devices that store information, data, and/or instructions that may be used by the processing circuitry 1102. The memory 1104 may store any suitable instructions, data, or information, including a computer program, software, an application including one or more of logic, rules, code, tables, and/or other instructions (collectively denoted computer program product 1104a) capable of being executed by the processing circuitry 1102 and utilized by the network node 1100. The memory 1104 may be used to store any calculations made by the processing circuitry 1102 and/or any data received via the communication interface 1106. In some embodiments, the processing circuitry 1102 and memory 1104 is integrated.

The communication interface 1106 is used in wired or wireless communication of signaling and/or data between a network node, access network, and/or UE. As illustrated, the communication interface 1106 comprises port(s)/terminal(s) 1116 to send and receive data, for example to and from a network over a wired connection. The communication interface 1106 also includes radio front-end circuitry 1118 that may be coupled to, or in certain embodiments a part of, the antenna 1110. Radio front-end circuitry 1118 comprises filters 1120 and amplifiers 1122. The radio front-end circuitry 1118 may be connected to an antenna 1110 and processing circuitry 1102. The radio front-end circuitry may be configured to condition signals communicated between antenna 1110 and processing circuitry 1102. The radio front-end circuitry 1118 may receive digital data that is to be sent out to other network nodes or UEs via a wireless connection. The radio frontend circuitry 1118 may convert the digital data into a radio signal having the appropriate channel and bandwidth parameters using a combination of filters 1120 and/or amplifiers 1122. The radio signal may then be transmitted via the antenna 1110. Similarly, when receiving data, the antenna 1110 may collect radio signals which are then converted into digital data by the radio front-end circuitry 1118. The digital data may be passed to the processing circuitry 1102. In other embodiments, the communication interface may comprise different components and/or different combinations of components.

In certain alternative embodiments, the network node 1100 does not include separate radio front-end circuitry 1118, instead, the processing circuitry 1102 includes radio front-end circuitry and is connected to the antenna 1110. Similarly, in some embodiments, all or some of the RF transceiver circuitry 1112 is part of the communication interface 1106. In still other embodiments, the communication interface 1106 includes one or more ports or terminals 1116, the radio frontend circuitry 1118, and the RF transceiver circuitry 1112, as part of a radio unit (not shown), and the communication interface 1106 communicates with the baseband processing circuitry 1114, which is part of a digital unit (not shown).

The antenna 1110 may include one or more antennas, or antenna arrays, configured to send and/or receive wireless signals. The antenna 1110 may be coupled to the radio front-end circuitry 1118 and may be any type of antenna capable of transmitting and receiving data and/or signals wirelessly. In certain embodiments, the antenna 1110 is separate from the network node 1100 and connectable to the network node 1100 through an interface or port.

The antenna 1110, communication interface 1106, and/or the processing circuitry 1102 may be configured to perform any receiving operations and/or certain obtaining operations described herein as being performed by the network node. Any information, data and/or signals may be received from a UE, another network node and/or any other network equipment. Similarly, the antenna 1110, the communication interface 1106, and/or the processing circuitry 1102 may be configured to perform any transmitting operations described herein as being performed by the network node. Any information, data and/or signals may be transmitted to a UE, another network node and/or any other network equipment.

The power source 1108 provides power to the various components of network node 1100 in a form suitable for the respective components (e.g., at a voltage and current level needed for each respective component). The power source 1108 may further comprise, or be coupled to, power management circuitry to supply the components of the network node 1100 with power for performing the functionality described herein. For example, the network node 1100 may be connectable to an external power source (e.g., the power grid, an electricity outlet) via an input circuitry or interface such as an electrical cable, whereby the external power source supplies power to power circuitry of the power source 1108. As a further example, the power source 1108 may comprise a source of power in the form of a battery or battery pack which is connected to, or integrated in, power circuitry. The battery may provide backup power should the external power source fail.

Embodiments of the network node 1100 may include additional components beyond those shown in Figure 11 for providing certain aspects of the network node’s functionality, including any of the functionality described herein and/or any functionality necessary to support the subject matter described herein. For example, the network node 1100 may include user interface equipment to allow input of information into the network node 1100 and to allow output of information from the network node 1100. This may allow a user to perform diagnostic, maintenance, repair, and other administrative functions for the network node 1100.

Figure 12 is a block diagram of a host 1200, which may be an embodiment of the host 1016 of Figure 10, in accordance with various aspects described herein. As used herein, the host 1200 may be or comprise various combinations hardware and/or software, including a standalone server, a blade server, a cloud-implemented server, a distributed server, a virtual machine, container, or processing resources in a server farm. The host 1200 may provide one or more services to one or more UEs and/or to other nodes or function in a communication network, such as SMO system 1020 shown in Figure 10.

The host 1200 includes processing circuitry 1202 that is operatively coupled via a bus 1204 to an input/output interface 1206, a network interface 1208, a power source 1210, and a memory 1212. Other components may be included in other embodiments. Features of these components may be similar to the components shown in Figure 11 , such that the above descriptions of those components are generally applicable to corresponding components of host 1200.

The memory 1212 may include one or more computer programs including one or more host application programs 1214 and data 1216, which may include user data, e.g., data generated by a UE for the host 1200 or data generated by the host 1200 for a UE. Embodiments of the host 1200 may utilize only a subset or all of the components shown. The host application programs 1214 may be implemented in a container-based architecture and may provide support for video codecs (e.g., Versatile Video Coding (VVC), High Efficiency Video Coding (HEVC), Advanced Video Coding (AVC), MPEG, VP9) and audio codecs (e.g., FEAC, Advanced Audio Coding (AAC), MPEG, G.711), including transcoding for multiple different classes, types, or implementations of UEs (e.g., handsets, desktop computers, wearable display systems, heads-up display systems). The host application programs 1214 may also provide for user authentication and licensing checks and may periodically report health, routes, and content availability to a central node, such as a device in or on the edge of a core network. Accordingly, the host 1200 may select and/or indicate a different host for over-the-top services for a UE. The host application programs 1214 may support various protocols, such as the HTTP Live Streaming (HLS) protocol, Real-Time Messaging Protocol (RTMP), Real-Time Streaming Protocol (RTSP), Dynamic Adaptive Streaming over HTTP (MPEG-DASH), etc.

In some embodiments, host 1200 can be configured to perform various operations of exemplary methods (e.g., procedures) for managing communication network resources based on end-user QoE, such as describe above in relation to Figure 9.

Figure 13 is a block diagram illustrating a virtualization environment 1300 in which functions implemented by some embodiments may be virtualized. In the present context, virtualizing means creating virtual versions of apparatuses or devices which may include virtualizing hardware platforms, storage devices and networking resources. As used herein, virtualization can be applied to any device described herein, or components thereof, and relates to an implementation in which at least a portion of the functionality is implemented as one or more virtual components. Some or all of the functions described herein may be implemented as virtual components executed by one or more virtual machines (VMs) implemented in one or more virtual environments 1300 hosted by one or more of hardware nodes, such as a hardware computing device that operates as a network node, UE, core network node, or host. Further, in embodiments in which the virtual node does not require radio connectivity (e.g., a core network node or host), then the node may be entirely virtualized.

Applications 1302 (which may alternatively be called software instances, virtual appliances, network functions, virtual nodes, virtual network functions, etc.) are run in the virtualization environment 1300 to implement some of the features, functions, and/or benefits of some of the embodiments disclosed herein. In some embodiments, one or more applications 1302 can be configured to perform various operations of exemplary methods e.g., procedures) for managing communication network resources based on end-user QoE, such as describe above in relation to Figure 9.

Hardware 1304 includes processing circuitry, memory that stores software and/or instructions (collectively denoted computer program product 1304a) executable by hardware processing circuitry, and/or other hardware devices as described herein, such as a network interface, input/output interface, and so forth. Software may be executed by the processing circuitry to instantiate one or more virtualization layers 1306 (also referred to as hypervisors or virtual machine monitors (VMMs)), provide VMs 1308a and 1308b (one or more of which may be generally referred to as VMs 1308), and/or perform any of the functions, features and/or benefits described in relation with some embodiments described herein. The virtualization layer 1306 may present a virtual operating platform that appears like networking hardware to the VMs 1308.

The VMs 1308 comprise virtual processing, virtual memory, virtual networking or interface and virtual storage, and may be run by a corresponding virtualization layer 1306. Different embodiments of the instance of a virtual appliance 1302 may be implemented on one or more of VMs 1308, and the implementations may be made in different ways. Virtualization of the hardware is in some contexts referred to as network function virtualization (NFV). NFV may be used to consolidate many network equipment types onto industry standard high volume server hardware, physical switches, and physical storage, which can be located in data centers, and customer premise equipment.

In the context of NFV, a VM 1308 may be a software implementation of a physical machine that runs programs as if they were executing on a physical, non-virtualized machine. Each of the VMs 1308, and that part of hardware 1304 that executes that VM, be it hardware dedicated to that VM and/or hardware shared by that VM with others of the VMs, forms separate virtual network elements. Still in the context of NFV, a virtual network function is responsible for handling specific network functions that run in one or more VMs 1308 on top of the hardware 1304 and corresponds to the application 1302.

Hardware 1304 may be implemented in a standalone network node with generic or specific components. Hardware 1304 may implement some functions via virtualization. Alternatively, hardware 1304 may be part of a larger cluster of hardware (e.g., in a data center) where many hardware nodes work together and are managed via management and orchestration 1310, which, among others, oversees lifecycle management of applications 1302. In some embodiments, hardware 1304 is coupled to one or more radio units that each include one or more transmitters and one or more receivers that may be coupled to one or more antennas. Radio units may communicate directly with other hardware nodes via one or more appropriate network interfaces and may be used in combination with the virtual components to provide a virtual node with radio capabilities, such as a radio access node or a base station. In some embodiments, some signaling can be provided with the use of a control system 1312 which may alternatively be used for communication between hardware nodes and radio units.

The foregoing merely illustrates the principles of the disclosure. Various modifications and alterations to the described embodiments will be apparent to those skilled in the art in view of the teachings herein. It will thus be appreciated that those skilled in the art will be able to devise numerous systems, arrangements, and procedures that, although not explicitly shown or described herein, embody the principles of the disclosure and can be thus within the spirit and scope of the disclosure. Various embodiments can be used together with one another, as well as interchangeably therewith, as should be understood by those having ordinary skill in the art.

The term unit, as used herein, can have conventional meaning in the field of electronics, electrical devices and/or electronic devices and can include, for example, electrical and/or electronic circuitry, devices, modules, processors, memories, logic solid state and/or discrete devices, computer programs or instructions for carrying out respective tasks, procedures, computations, outputs, and/or displaying functions, and so on, as such as those that are described herein.

Any appropriate steps, methods, features, functions, or benefits disclosed herein may be performed through one or more functional units or modules of one or more virtual apparatuses. Each virtual apparatus may comprise a number of these functional units. These functional units may be implemented via processing circuitry, which may include one or more microprocessor or microcontrollers, as well as other digital hardware, which may include Digital Signal Processor (DSPs), special-purpose digital logic, and the like. The processing circuitry may be configured to execute program code stored in memory, which may include one or several types of memory such as Read Only Memory (ROM), Random Access Memory (RAM), cache memory, flash memory devices, optical storage devices, etc. Program code stored in memory includes program instructions for executing one or more telecommunications and/or data communications protocols as well as instructions for carrying out one or more of the techniques described herein. In some implementations, the processing circuitry may be used to cause the respective functional unit to perform corresponding functions according one or more embodiments of the present disclosure.

As described herein, device and/or apparatus can be represented by a semiconductor chip, a chipset, or a (hardware) module comprising such chip or chipset; this, however, does not exclude the possibility that a functionality of a device or apparatus, instead of being hardware implemented, be implemented as a software module such as a computer program or a computer program product comprising executable software code portions for execution or being run on a processor. Furthermore, functionality of a device or apparatus can be implemented by any combination of hardware and software. A device or apparatus can also be regarded as an assembly of multiple devices and/or apparatuses, whether functionally in cooperation with or independently of each other. Moreover, devices and apparatuses can be implemented in a distributed fashion throughout a system, so long as the functionality of the device or apparatus is preserved. Such and similar principles are considered as known to a skilled person.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. It will be further understood that terms used herein should be interpreted as having a meaning that is consistent with their meaning in the context of this specification and the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

In addition, certain terms used in the present disclosure, including the specification and drawings, can be used synonymously in certain instances (e.g., “data” and “information”). It should be understood, that although these terms (and/or other terms that can be synonymous to one another) can be used synonymously herein, there can be instances when such words can be intended to not be used synonymously.

Claims

1. A computer-implemented method for managing communication network resources based on end-user quality of experience, QoE, the method comprising: collecting (910) performance-related data from a radio access network, RAN, and a core network, CN, comprising the communication network; correlating (930) at least the collected performance-related data into a plurality of records corresponding to a plurality of user data sessions with the communication network; training (940) one or more artificial intelligence/machine learning, AI/ML, models based on the correlated data records, wherein each AI/ML model maps between network quality-of-service, QoS, metrics and service-specific QoE metrics; and configuring (960) communication network resources to carry one or more second user data sessions, based on the trained AI/ML model and on QoE requirements of respective services comprising the second user data sessions.

2. The method of claim 1, wherein the performance-related data includes one or more of the following: trace data for respective cells provided by RAN nodes; performance management, PM, counter data associated with the RAN nodes; user plane, UP, event information associated with the CN; and control plane, CP, event information associated with the CN.

3. The method of claim 2, wherein the UP event information includes information that identifies one or more of the following: respective user data sessions; user subscriptions associated with the respective user data sessions; user equipment, UEs, associated with the respective user data sessions; services associated with the respective user data sessions; and network QoS metrics associated with at least one of the following: the respective user data sessions, and the services associated with the respective user data sessions.

4. The method of any of claims 1-3, wherein each correlated data record includes one or more of the following associated with the corresponding user data session: an identifier of a service; an identifier of user network subscription; identifier of a user equipment, UE; location of the UE; mobility history of the UE; radio access technology, RAT, being used in the RAN; and an identifier corresponding to a plurality of network QoS characteristics.

5. The method of any of claims 1-4, wherein correlating (930) at least the collected performance -related data comprises determining (931), from the collected performance-related data, the following network QoS metrics associated with each user data session: one or more RAN QoS metrics, and one or more CN QoS metrics.

6. The method of claim 5, wherein: the RAN QoS metrics associated with each user data session include one or more of the following: RAN resources used, serving cell load, mobility events between serving cells, and serving and neighbor cell radio measurements; and the CN QoS metrics associated with each user data session include one or more of the following: packet delay, packet delay jitter, packet loss, and priority level.

7. The method of any of claims 5-6, wherein: the plurality of user data sessions are associated with a corresponding plurality of services; the method further comprises collecting (920) service-specific QoE data for the plurality of user data sessions; and correlating (930) at least the collected performance-related data further comprises correlating (932) the collected service- specific QoE data into the plurality of data records.

8. The method of claim 7, where the service-specific QoE data includes one or more of the following: opinion score, service availability, bitrate statistics, playback errors, rebujfering, startup time, startup failures, and duration of user data session.

9. The method of any of claims 5-8, wherein training (940) each AI/ML model based on the correlated data records is performed using the network QoS metrics as input features and the service- specific QoE data as output labels.

10. The method of any of claims 1-9, wherein: the method further comprises determining (950) a plurality of service- specific policies for network resource management, based on the one or more trained AI/ML models and on QoE requirements of a respective plurality of services; and each service-specific policy includes corresponding network QoS requirements.

11. The method of claim 10, wherein configuring (960) communication network resources to carry one or more second user data sessions comprises enforcing (961) service-specific policies for the respective second user data sessions.

12. The method of claim 11, wherein enforcing (961) service-specific policies for the respective second user data sessions comprises, for each second user data session: determining (961a) network QoS requirements for the second user data session based on the one or more trained AI/ML models and on QoE requirements of a service comprising the second user data session; selecting (961b) one of the service-specific policies that includes network QoS requirements that meet the network QoS requirements for the second user data session; and configuring (961c) communication network resources for the second user data session in accordance with the network QoS requirements included in the selected servicespecific policy.

13. The method of claim 12, further comprising collecting (970) one or more of the following feedback for the respective second user data sessions: observed network QoS metrics associated with the configured communication resources, and observed QoE for the respective services.

14. The method of claim 13, further comprising retraining (980) the one or more AI/ML models based on the collected feedback.

15. The method of claim 13, wherein: a first service-specific policy is selected for a first subset of the second user data sessions; a second service-specific policy is selected for a second subset of the second user data sessions; and the method further comprises performing (990) reinforcement learning, RL, for policy selection based on the collected feedback for the second user data sessions.

16. The method of any of claims 1-15, wherein configuring (960) communication network resources to carry one or more second user data sessions comprises changing (965) one or more network resources or settings for each second user data session based on the respective QoE requirements of the services comprising the second user data sessions.

17. The method of claim 16, wherein the one or more network resources or settings include one or more of the following: serving cell; resource allocation within serving cell; serving radio access technology, RAT; serving RAN node; and QoS settings.

18. The method of claim 17, wherein configuring (960) communication network resources to carry one or more second user data sessions further comprises, while a plurality of user data sessions are active, detecting (962) a change in network QoS metrics indicating a degradation in communication network performance.

19. The method of claim 18, wherein: configuring (960) communication network resources to carry one or more second user data sessions further comprises: using the one or more AI/ML models, predicting (963) QoE for the respective services comprising the active user data sessions based on the detected change in network QoS metrics; and refraining (966) from changing network resources or settings for one or more third user data sessions comprising services having predicted QoE that meet corresponding QoE requirements; and the one or more second user data sessions, whose network resources or settings are changed, comprise services having predicted QoE that do not meet corresponding QoE requirements.

20. The method of claim 19, further comprising, using the one or more AI/ML models, predicting (964) further QoE for the respective services comprising the second user data sessions based on a further change in network QoS metrics associated with the changed network resources or settings, wherein the predicted further QoE meet the corresponding QoE requirements.

2L The method of claim 19, wherein: the one or more second user data sessions are associated with one or more of the following: lower-priority user subscription profiles, and lower-mobility users; and configuring (960) communication network resources to carry one or more second user data sessions further comprises refraining (967) from changing network resources or settings for one or more third user data sessions associated with one or more of the following: higher-priority user subscription profiles, and higher-mobility users.

22. The method of any of claims 1-21, wherein: configuring (960) communication network resources is performed by a non-real-time RAN intelligent controller of a service management and orchestration (SMO) system of the communication network; and the collecting (910), correlating (930), and training (940) operations are performed by one of the following: the SMO system, or a cloud computing environment external to the SMO system.

23. The method of any of claims 1-22, wherein the RAN is an Open RAN.

24. A network management system (400, 500, 600, 700, 800, 1016, 1018, 1020) configured to manage resources of a communication network (100, 530, 1002) based on end-user quality of experience, QoE, the network management system comprising: communication interface circuitry (1106, 1304) configured to communicate a radio access network, RAN (199, 510, 1004) and a core network, CN (198, 520, 1006) comprising the communication network; and processing circuitry (1102, 1304) operatively coupled to the communication interface circuitry, whereby the processing circuitry and the communication interface circuitry are configured to: collect performance-related data from the RAN and the CN; correlate at least the collected performance-related data into a plurality of records corresponding to a plurality of user data sessions with the communication network; train one or more artificial intelligence/machine learning, AI/ML, models based on the correlated data records, wherein each AI/ML model maps between network quality-of-service, QoS, metrics and service-specific QoE metrics; and configure communication network resources to carry one or more second user data sessions, based on the trained AI/ML model and on QoE requirements of respective services comprising the second user data sessions.

25. The network management system of claim 24, wherein the processing circuitry and the communication interface circuitry are further configured to perform operations corresponding to any of claims 2-23.

26. A network management system (400, 500, 600, 700, 800, 1016, 1018, 1020) configured to manage resources of a communication network (100, 530, 1002) based on end-user quality of experience, QoE, the network management system being further configured to: collect performance-related data from a radio access network, RAN (199, 510, 1004) and a core network, CN (198, 520, 1006) comprising the communication network; correlate at least the collected performance-related data into a plurality of records corresponding to a plurality of user data sessions with the communication network; train one or more artificial intelligence/machine learning, AI/ML, models based on the correlated data records, wherein each AI/ML model maps between network quality-of-service, QoS, metrics and service-specific QoE metrics; and configure communication network resources to carry one or more second user data sessions, based on the trained AI/ML model and on QoE requirements of respective services comprising the second user data sessions.

27. The network management system of claim 26, being further configured to perform operations corresponding to any of the methods of claims 2-23.

28. The network management system of claim 26, comprising: an analytics function configured to collect the performance -related data and to correlate at least the collected performance -related data; a learning system configured to train the one or more AI/ML models based on the correlated data records; and a policy enforcement function and a traffic steering function, collectively arranged to configure the communication network resources to carry one or more second user data sessions.

29. A non-transitory, computer-readable medium (1104, 1304) storing computer-executable instructions that, when executed by processing circuitry (1102, 1304) associated with a network management system (400, 500, 600, 700, 800, 1016, 1018, 1020) configured to manage resources of a communication network (100, 530, 1002) based on end-user quality of experience, QoE, configure the network management system to perform operations corresponding to any of the methods of claims 1-23.

30. A computer program product (1104a, 1304a) comprising computer-executable instructions that, when executed by processing circuitry (1102, 1304) associated with a network management system (400, 500, 600, 700, 800, 1016, 1018, 1020) configured to manage resources of a communication network (100, 530, 1002) based on end-user quality of experience, QoE, configure the network management system to perform operations corresponding to any of the methods of claims 1-23.