IP multimedia subsystem and session policy optimization method

Abstract

The invention discloses an IP multimedia subsystem and a session policy optimization method. The IP multimedia subsystem comprises a policy optimization module and an intelligent analysis and prediction module. The strategy optimization module is configured to send a strategy request to the PCF before a session is established, and request the PCF to provide an initial strategy parameter; and in the session proceeding process, requesting the PCF to issue a new policy scheme. The intelligent analysis and prediction module is configured to obtain current and historical network state data from the NWDAF; predicting based on the network state data to obtain network prediction data; and the policy optimization module is used for making a decision on whether policy updating is needed or not based on the network prediction data and outputting a policy updating suggestion to the policy optimization module, and the policy optimization module is used for determining whether to request the PCF to issue a new policy scheme or not according to the policy updating suggestion and requesting the PCF to issue the new policy scheme according to the policy updating suggestion when determining that the PCF is required to issue the new policy scheme.

Description

Technical Field

[0001] This application relates to the field of communication technologies, and particularly to an IP multimedia subsystem and a session policy optimization method. Background Art

[0002] The IP Multimedia Subsystem (IMS) is an IP-based communication architecture, initially standardized by 3GPP (3rd Generation Partnership Project) for providing multimedia services in mobile and fixed networks. The core of IMS technology lies in its hierarchical architecture and standardized interfaces, which enable various multimedia services to be managed and provided through a unified platform. Currently, IMS has become the standard platform for telecom operators to provide voice, video, and other multimedia services, and is widely used in services such as VoLTE (Voice over LTE) and VoWiFi (Voice over Wi-Fi).

[0003] With the deployment of 5G networks, the importance of IMS has been further highlighted. 5G is not just about speed improvement, but also a brand-new network ecosystem that supports diverse application scenarios. As a core part of the 5G network, IMS will play a key role in supporting ultra-low latency communication, massive machine-type communication, and enhanced mobile broadband services.

[0004] However, the current IMS multimedia telephone service has the following deficiencies:

[0005] 1. Insufficient intelligence: Existing solutions mostly rely on predefined policies and manual configuration (static policies), lacking the ability of dynamic optimization and intelligent decision-making based on real-time data, and it is difficult to cope with complex and changing network environments and user requirements.

[0006] 2. Limited data analysis ability: Although NWDAF is introduced for data analysis, in the specific scenarios of multimedia services, the depth and breadth of data analysis are limited, and it is difficult to fully explore potential patterns in user behavior and / or network operation, resulting in unsatisfactory optimization effects.

[0007] 3. Low system integration: In the integration process of the 5G core network and IMS in existing solutions, there is a lack of unified system architecture design, resulting in low collaborative work efficiency among functional modules and it is difficult to achieve overall performance improvement.

[0008] 4. Unoptimized resource management: In multimedia services, the allocation and management of network resources are often not fine enough, and it is unable to dynamically adjust according to real-time requirements and network conditions, resulting in resource waste or service quality degradation. Summary of the Invention

[0009] The purpose of this application is to provide an IP multimedia subsystem and a session policy optimization method, which can solve at least one technical problem existing in the background technology.

[0010] To achieve the above object, this application provides an IP multimedia subsystem, including a policy optimization module and an intelligent analysis and prediction module.

[0011] The policy optimization module is configured to:

[0012] Before session establishment, send a policy request to the PCF, requesting the PCF to provide initial policy parameters;

[0013] During the session, request the PCF to issue a new policy plan;

[0014] The intelligent analysis and prediction module is configured to:

[0015] Obtain current and historical network status data from the NWDAF;

[0016] Based on the network status data, predict network prediction data;

[0017] Based on the network prediction data, make a decision on whether policy update is required and output a policy update suggestion to the policy optimization module. The policy optimization module determines whether to request the PCF to issue a new policy plan according to the policy update suggestion and requests the PCF to issue a new policy plan according to the policy update suggestion when it is determined to request the PCF to issue a new policy plan.

[0018] Optionally, the policy optimization module includes a policy request interface, a policy update interface, and a policy feedback interface;

[0019] The policy request interface is configured to: before session establishment, send a policy request to the PCF, requesting the PCF to provide the initial policy parameters;

[0020] The policy update interface is configured to: during the session, request the PCF to issue a new policy plan;

[0021] The policy feedback interface is configured to: after policy execution is completed, feedback the policy execution situation to the PCF;

[0022] The intelligent analysis and prediction module includes a data acquisition interface and a decision result interface;

[0023] The data acquisition interface is configured to obtain network status data from the NWDAF;

[0024] The result output interface is configured to send the policy update suggestion to the policy optimization module.

[0025] Optionally, the network status analysis data includes: bandwidth utilization rate, historical data traffic, latency, and packet loss rate;

[0026] The network prediction data includes: bandwidth utilization rate prediction data, latency prediction data, and packet loss rate prediction data;

[0027] The intelligent analysis and prediction module is further configured to: obtain congestion probability prediction data from the NWDAF;

[0028] The intelligent analysis and prediction module makes a decision on whether policy update is needed based on the network prediction data and the congestion probability prediction data and outputs the policy update recommendation.

[0029] Optionally, the intelligent analysis and prediction module is further configured to: obtain the user's current policy, the user's current QoE, and the user priority;

[0030] The intelligent analysis and prediction module makes a decision on whether policy update is needed based on the network prediction data, the congestion probability prediction data, the user's current policy, the user's current QoE, and the user priority and outputs the policy update recommendation.

[0031] Optionally, the IP multimedia subsystem includes a CSCF, an MRF, and an AS. The policy optimization module dynamically adjusts the session policy based on the signaling control of the CSCF and the media processing information of the MRF. The intelligent analysis and prediction module obtains service data from the AS and mines the user behavior pattern and predicts the change of service demand based on the service data.

[0032] To achieve the above object, the present application further provides a session policy optimization method based on the above IP multimedia subsystem, including the following steps:

[0033] S1, the IP multimedia subsystem receives a session establishment request from a UE;

[0034] S2, the policy optimization module requests the PCF to provide the initial policy parameters and receives the initial policy parameters returned by the PCF for configuration so that the IP multimedia subsystem executes the initial policy;

[0035] S3, the intelligent analysis and prediction module obtains the network status data from the NWDAF and predicts the network prediction data based on the network status data;

[0036] S4, the intelligent analysis and prediction module makes a decision on whether policy update is needed based on the network prediction data and outputs the policy update recommendation to the policy optimization module;

[0037] S5. The policy optimization module determines whether to request the PCF to issue a new policy plan according to the policy update suggestion. If the determination result is yes, it proceeds to step S6;

[0038] S6. The policy optimization module requests the PCF to issue a new policy plan based on the policy update suggestion and updates the configuration parameters of the IP multimedia subsystem based on the new policy plan;

[0039] S7. The IP multimedia subsystem executes the new policy.

[0040] Optionally, after step S7, the method further includes:

[0041] S8. After the current session ends, the policy optimization module sends the policy execution status to the PCF;

[0042] In step S5, if the determination result is no, it proceeds to step S8.

[0043] Optionally, the intelligent analysis and prediction module makes a decision on whether policy update is needed based on the network prediction data and outputs the policy update suggestion to the policy optimization module. The intelligent analysis and prediction module includes:

[0044] Compare the network prediction data with the prediction data threshold, and make a decision on whether policy update is needed according to the comparison result and output the policy update suggestion to the policy optimization module.

[0045] Optionally, the input parameters of the intelligent analysis and prediction module further include: NWDAF prediction data, user's current policy, user's current QoE, and user priority;

[0046] The intelligent analysis and prediction module makes a decision on whether policy update is needed based on the network prediction data, the NWDAF prediction data, the user's current policy, the user's current QoE, and the user priority and outputs the policy update suggestion to the policy optimization module.

[0047] Optionally, the method further includes:

[0048] The NWDAF predicts the prediction data of the user behavior pattern based on the historical data and real-time data of the user behavior pattern;

[0049] The intelligent analysis and prediction module makes a decision on whether policy optimization is needed based on the prediction data of the user behavior pattern obtained from the NWDAF and outputs the policy optimization suggestion to the policy optimization module;

[0050] The policy optimization module requests the PCF to issue a new policy plan based on the policy optimization suggestion and updates the configuration parameters of the IP multimedia subsystem based on the new policy plan.

[0051] Compared with the traditional IP multimedia subsystem, the IP multimedia subsystem of the embodiment of the present application additionally includes a policy optimization module and an intelligent analysis and prediction module. The policy optimization module can request the PCF to provide initial policy parameters and issue a new policy plan when needed. The intelligent analysis and prediction module can obtain current and historical network status data from the NWDAF and predict network prediction data based on this, and then make a decision on whether policy update is needed and output a policy update suggestion to the policy optimization module, so that the policy optimization module can determine whether to request the PCF to issue a new policy plan based on the policy update suggestion and request the PCF to issue a new policy plan based on the policy update suggestion. It can be seen that in the embodiment of the present application, the intelligent analysis and prediction module can be used as an enhanced function of the NWDAF, with the ability to directly generate optimization suggestions from network data, shortening the policy decision link between the IP multimedia subsystem and the NWDAF and improving the policy adjustment efficiency. The embodiment of the present application can realize the dynamic optimization and intelligent decision-making of multimedia telephone services, improve the adaptive ability of the services and the user experience. Moreover, based on the setting of the intelligent analysis and prediction module docked with the NWDAF, it is convenient to comprehensively analyze the key data in the IMS multimedia telephone service, which is beneficial to mining potential patterns and trends and providing a basis for service optimization. In addition, the embodiment of the present application designs a unified system architecture to closely combine the 5G core network with the IP multimedia subsystem, which is beneficial to the efficient cooperation between relevant functional modules and improves the performance and reliability of the overall system. In addition, through intelligent resource scheduling, according to the real-time network conditions, the network resources required for multimedia telephone services can be dynamically adjusted, improving resource utilization and service quality. Description of the Drawings

[0052] Figure 1 It is a schematic block diagram of an IP multimedia subsystem and a 5G core network.

[0053] Figure 2 It is a schematic block diagram of the policy optimization module and the PCF of the embodiment of the present application.

[0054] Figure 3 It is a schematic block diagram of the NWDAF, the intelligent analysis and prediction module and the PCF of the embodiment of the present application.

[0055] Figure 4 It is a flowchart of the session policy optimization method based on the IP multimedia subsystem of the embodiment of the present application. Detailed Embodiments

[0056] To elaborate in detail on the technical content, structural features, achieved objectives, and effects of this application, the following provides a detailed explanation in conjunction with the embodiments and accompanied by the drawings.

[0057] To facilitate the understanding of this application, the relevant terms that appear in this application are explained as follows:

[0058] IMS (IP Multimedia Subsystem): An IP-based communication architecture that provides multimedia services such as voice, video, instant messaging, etc.

[0059] CSCF (Call Session Control Function): The session control function within the IP multimedia subsystem, which is the core of the entire IMS network.

[0060] MRF (Media Resource Function): The media resource function in IMS, which is the component responsible for processing media streams and can control and optimize media streams.

[0061] AS (Application Server): The application server in IMS that can interact with user devices to provide services.

[0062] PCF (Policy Control Function): The policy control function in the 5G core network, which is responsible for the management of network resources and the execution of user policies.

[0063] NWDAF (Network Data Analytics Function): The network data analytics function in the 5G core network that provides network optimization suggestions through data analysis.

[0064] VoIP (Voice over IP): A technology for voice communication over the Internet protocol.

[0065] 5G core network: The core network of the fifth-generation mobile communication technology, which supports high-speed, low-latency large-scale communication connections.

[0066] This application can be applied at least to the call services and VoIP services of smartphones and mobile communication devices. It can enhance the interactive experience of users during calls, facilitate the provision of real-time suggestions and auxiliary services. Based on integrating the technical solution of this application into the VoIP system, intelligent functions can be added to traditional voice calls, enhancing the added value of the services.

[0067] Please refer to Figures 1 to 3, an embodiment of the present application discloses an IP Multimedia Subsystem (IMS), including a policy optimization module 10 and an intelligent analysis and prediction module 20.

[0068] The policy optimization module 10 is configured to:

[0069] Before session establishment, send a policy request to the PCF 30, requesting the PCF 30 to provide initial policy parameters;

[0070] During the session, request the PCF 30 to issue a new policy plan.

[0071] It should be noted that the policy optimization module 10 can be integrated inside the original IP Multimedia Subsystem or exist as an extended component of the original IP Multimedia Subsystem.

[0072] Specifically, the policy optimization module 10 in the IP Multimedia Subsystem is responsible for combining the policy rules issued by the PCF 30 with the business logic of the IP Multimedia Subsystem and performing the following functions according to the session context and dynamic network conditions:

[0073] Call path optimization: Under multi-access and multi-bearer conditions, select the optimal media path and routing according to the PCF 30 policy and user priority.

[0074] Dynamic adjustment of resource allocation strategy: During session establishment or progress, dynamically adjust bandwidth allocation, codec parameters, transcoding strategy, and media stream transmission priority according to the PCF 30 policy.

[0075] QoS (Quality of Service) and QoE (Quality of Experience) guarantee: Dynamically allocate resources according to user priority and current network load conditions to ensure high-priority service quality for specific users (such as VIP users or services with low latency requirements).

[0076] The intelligent analysis and prediction module 20 is configured to:

[0077] Obtain current and historical network status data from the NWDAF 40;

[0078] Based on the network status data, obtain network prediction data;

[0079] Based on the network prediction data, make a decision on whether policy update is needed and output a policy update recommendation to the policy optimization module 10. The policy optimization module 10 determines whether to request the PCF 30 to issue a new policy plan according to the policy update recommendation and requests the PCF 30 to issue a new policy plan according to the policy update recommendation when it is determined to request the PCF 30 to issue a new policy plan.

[0080] It should be noted that the intelligent analysis and prediction module 20 can be integrated inside the original IP multimedia subsystem or exist as an extended component of the original IP multimedia subsystem.

[0081] It should be noted that the policy update suggestion can include not updating, that is, keeping the original policy.

[0082] It should be noted that the data output by the intelligent analysis and prediction module 20 to the policy optimization module 10 can also include: user ID. Additionally, it can also include data such as QoE change prediction and confidence level.

[0083] Specifically, the intelligent analysis and prediction module 20 can predict the future network status and user requirements both before and during the session. For example, it predicts that network congestion may occur within the next 5 seconds and thus issues an early warning to the policy optimization module 10 in advance. The policy update suggestions put forward by the intelligent analysis and prediction module 20 to the policy optimization module 10 can include reducing the bit rate, switching the media route, increasing redundant coding, etc., to ensure the call quality.

[0084] Compared with the traditional IP multimedia subsystem, the IP multimedia subsystem of the embodiment of the present application additionally includes a policy optimization module 10 and an intelligent analysis and prediction module 20. The policy optimization module 10 can request the PCF 30 to provide initial policy parameters and issue a new policy plan when needed. The intelligent analysis and prediction module 20 can obtain current and historical network status data from the NWDAF 40 and predict network prediction data based on this, and then can make a decision on whether policy update is needed and output a policy update suggestion to the policy optimization module 10, so that the policy optimization module 10 can determine whether to request the PCF 30 to issue a new policy plan based on the policy update suggestion and request the PCF 30 to issue a new policy plan based on the policy update suggestion. It can be seen that in the embodiment of the present application, the intelligent analysis and prediction module 20 can serve as an enhanced function of the NWDAF 40, have the ability to directly generate optimization suggestions from network data, shorten the policy decision link between the IP multimedia subsystem and the NWDAF 40, and improve the policy adjustment efficiency. The embodiment of the present application can realize the dynamic optimization and intelligent decision-making of multimedia telephone services, improve the adaptive ability and user experience of the services. Moreover, based on the setting of the intelligent analysis and prediction module 20 docked with the NWDAF 40, it is convenient to comprehensively analyze the key data in the IMS multimedia telephone service, which is beneficial to mining potential patterns and trends and providing a basis for service optimization. In addition, the embodiment of the present application designs a unified system architecture to closely combine the 5G core network with the IP multimedia subsystem, which is beneficial to the efficient cooperation between relevant functional modules and improves the performance and reliability of the overall system. In addition, through intelligent resource scheduling, according to the real-time network conditions, the network resources required for multimedia telephone services can be dynamically adjusted, improving resource utilization and service quality.

[0085] Through the establishment of the policy optimization module 10 and the intelligent analysis and prediction module 20, it is beneficial for the IP multimedia subsystem to achieve dynamic and multi-dimensional policy optimization in each stage before, during, and after a call.

[0086] Specifically, the policy optimization module 10 includes a policy request interface SRI, a policy update interface, and a policy feedback interface.

[0087] The policy request interface SRI is configured to: before session establishment, send a policy request to the PCF 30 (the request information may include user ID, service type, expected bandwidth requirement, etc.) to request the PCF 30 to provide initial policy parameters.

[0088] The policy update interface SUI is configured to: during the session, request the PCF 30 to issue a new policy plan.

[0089] The Policy Feedback Interface SFI is configured to: after the policy execution is completed, feedback the policy execution situation to the PCF 30, and provide reference data for the subsequent policy optimization of the PCF 30. In addition, after the policy execution is completed, the Policy Optimization Module 10 also feedbacks the end-user experience metrics to the PCF 30 through the Policy Feedback Interface SFI, so as to better provide reference data for the subsequent policy optimization of the PCF 30.

[0090] Through the setting of the above interfaces, the direct data interaction between the Policy Optimization Module 10 and the PCF 30 is ensured, which is conducive to the efficient cooperation between the Policy Optimization Module 10 and the PCF 30.

[0091] The Intelligent Analysis and Prediction Module 20 includes a Data Acquisition Interface DGI and a Result Output Interface PRI. The Data Acquisition Interface DGI is configured to obtain network status data from the NWDAF 40. The Result Output Interface PRI is configured to send policy update suggestions to the Policy Optimization Module 10.

[0092] Through the setting of the above interfaces, the direct connection between the Intelligent Analysis and Prediction Module 20 and the NWDAF 40 is ensured, which is conducive to the efficient cooperation between the Intelligent Analysis and Prediction Module 20 and the NWDAF 40.

[0093] In some embodiments, the network status analysis data at least includes: bandwidth utilization rate, data traffic, latency (link latency), and packet loss rate. Parameters such as bandwidth utilization rate, historical data traffic, latency, and packet loss rate are provided as basic inputs to the prediction model of the Intelligent Analysis and Prediction Module 20. Specifically, machine learning (such as Transformer), statistical analysis, or deep learning algorithms can be used to predict the future network status.

[0094] Specifically, the bandwidth utilization rate involves the current available network bandwidth and the bandwidth occupancy trend. The historical data traffic can include the traffic growth trend in a certain area or a certain time period. Latency refers to the round-trip time of data packets on a certain link, which affects the quality of services such as VoIP and video calls. The packet loss rate refers to the packet loss situation within a certain period of time, which affects data reliability and user experience.

[0095] Specifically, after obtaining data from the NWDAF 40, the Intelligent Analysis and Prediction Module 20 will perform feature extraction on the data to screen out the data related to network quality.

[0096] The network prediction data at least includes: bandwidth utilization rate prediction data, latency prediction data, and packet loss rate prediction data.

[0097] Specifically, the bandwidth utilization prediction data can be the congestion risk predicted for a future time period or a specific area to anticipate network resource requirements. The latency prediction data can be the latency conditions predicted 5 seconds or 10 seconds in the future based on data traffic, network topology changes, etc. to optimize media stream routing. The packet loss rate prediction data can be the data obtained through historical trend analysis to give an early warning before the data link quality deteriorates and optimize the QoS policy or increase the redundancy mechanism in advance.

[0098] In some embodiments, the intelligent analysis and prediction module 20 is further configured to: obtain congestion probability prediction data from the NWDAF 40. The intelligent analysis and prediction module 20 makes a decision on whether a policy update is needed based on the network prediction data and the congestion probability prediction data and outputs a policy update recommendation.

[0099] Specifically, the intelligent analysis and prediction module 20 is further configured to: obtain the user's current policy, the user's current QoE, and the user priority. The intelligent analysis and prediction module 20 makes a decision on whether a policy update is needed based on the network prediction data, the congestion probability prediction data, the user's current policy, the user's current QoE, and the user priority and outputs a policy update recommendation.

[0100] By introducing a policy optimization mechanism for multiple users and multiple levels of priorities, and by intelligently analyzing user behavior and network load to design differentiated policies in a hierarchical manner, the needs of different users (such as VIP users and ordinary users) can be met.

[0101] Specifically, if the intelligent analysis and prediction module 20 predicts network degradation and the user experience is already low, the policy optimization module 10 preferentially requests the PCF 30 to increase the policy priority or reduce the bit rate; if there is only slight degradation, the bit rate is reduced in advance to prevent problems before they occur.

[0102] In some embodiments, the input data of the intelligent analysis and prediction module 20 may further include at least any one of user behavior characteristics, user location distribution, and historical communication records.

[0103] In the case where the input data of the intelligent analysis and prediction module 20 includes user behavior characteristics, the QoS policy can be optimized by analyzing user service preferences (such as high-bandwidth applications, VoIP calls, video traffic), the communication patterns of users can be predicted. For example, some users need higher bandwidth during specific periods, so intelligent adjustments (such as preallocating network resources or optimizing routing) can be made when high load is predicted. Different types of users' bandwidth requirements can be combined with AI analysis to enhance the personalized service ability.

[0104] When the input data of the intelligent analysis and prediction module 20 includes the user location distribution, it is possible to identify user-dense areas, predict the load fluctuations in specific areas (such as large event venues, airports, commercial centers), combine the service capabilities of base stations / cells, predict and optimize the load balancing strategy, reduce the probability of congestion, and analyze the mobility patterns through historical location data (such as commuting peak traffic) to adjust the network resource allocation in advance.

[0105] When the input data of the intelligent analysis and prediction module 20 includes historical communication records, the historical communication records can be used as a machine learning training data set to help predict future network states (such as congestion may occur at a certain area at 5 pm every day), identify periodic problems (such as network jitter during specific time periods, overload conditions of specific base stations), provide a basis for long-term optimization, and combine historical fault data to improve the adaptive optimization ability of network maintenance and reduce manual intervention.

[0106] In some embodiments, the IP multimedia subsystem includes a CSCF 50, an MRF 60, and an AS 70. The policy optimization module 10 dynamically adjusts the session policy based on the signaling control of the CSCF 50 and the media processing information of the MRF 60. The intelligent analysis and prediction module 20 obtains service data from the AS 70 and mines user behavior patterns and predicts changes in service requirements based on the service data.

[0107] Please combine Figures 1 to 4 In addition, an embodiment of the present application also discloses a session policy optimization method based on the above IP multimedia subsystem, including the following steps:

[0108] S1, the IP multimedia subsystem receives a session establishment request from a UE (user equipment).

[0109] S2, the policy optimization module 10 requests the PCF 30 to provide initial policy parameters and receives the initial policy parameters returned by the PCF 30 for configuration to enable the IP multimedia subsystem to execute the initial policy.

[0110] S3, the intelligent analysis and prediction module 20 obtains network status data from the NWDAF 40 and predicts network prediction data based on the network status data.

[0111] S4, the intelligent analysis and prediction module 20 makes a decision on whether policy update is required based on the network prediction data and outputs a policy update recommendation to the policy optimization module 10.

[0112] S5, the policy optimization module 10 determines whether to request the PCF 30 to issue a new policy plan according to the policy update recommendation. If the determination result is yes, it proceeds to step S6.

[0113] S6, the policy optimization module 10 requests the PCF 30 to issue a new policy solution based on the policy update recommendation and updates the configuration parameters of the IP multimedia subsystem based on the new policy solution.

[0114] S7, the IP multimedia subsystem executes the new policy.

[0115] The embodiments of the present application can achieve the dynamic optimization and intelligent decision-making of multimedia phone services, improve the adaptive ability of the services and the user experience, can comprehensively analyze the key data in the IMS multimedia phone service, is conducive to mining potential patterns and trends, provides a basis for service optimization, and the 5G core network is closely combined with the IP multimedia subsystem, which is conducive to the efficient collaboration between relevant functional modules and improves the performance and reliability of the overall system. In addition, through intelligent resource scheduling, according to the real-time network conditions, the network resources required for multimedia phone services can be dynamically adjusted, improving resource utilization and service quality.

[0116] In some embodiments, after step S7, the method further includes:

[0117] S8, after the current session ends, the policy optimization module 10 sends the policy execution status to the PCF 30;

[0118] In step S5, if the determination result is negative, then step S8 is entered.

[0119] Since the policy optimization module 10 will feedback the policy execution status to the PCF 30, it can provide reference data for the subsequent policy optimization of the PCF 30.

[0120] In some embodiments, the intelligent analysis and prediction module 20 makes a decision on whether a policy update is needed based on network prediction data and outputs a policy update recommendation to the policy optimization module 10, including:

[0121] Compare the network prediction data with the prediction data threshold, and make a decision on whether a policy update is needed according to the comparison result and output a policy update recommendation to the policy optimization module 10.

[0122] In some embodiments, the input parameters of the intelligent analysis and prediction module 20 further include: NWDAF 40 prediction data, the user's current policy, the user's current QoE, and the user priority. The intelligent analysis and prediction module 20 makes a decision on whether a policy update is needed based on the network prediction data, NWDAF prediction data, the user's current policy, the user's current QoE, and the user priority and outputs a policy update recommendation to the policy optimization module 10.

[0123] Specifically, the NWDAF prediction data, i.e., the data predicted by the NWDAF 40, is directly provided to the intelligent analysis and prediction module 20 as the basis for analysis and decision-making. The NWDAF prediction data may include congestion probability prediction data. The intelligent analysis and prediction module 20 may compare the congestion probability prediction data with the congestion probability occurrence probability threshold as the basis for decision-making.

[0124] In some embodiments, the NWDAF 40 predicts the predicted data of the user behavior pattern based on the historical data and real-time data of the user behavior pattern. The intelligent analysis and prediction module 20 makes a decision on whether policy optimization is required based on the predicted data of the user behavior pattern obtained from the NWDAF 40 and outputs a policy optimization recommendation to the policy optimization module 10. The policy optimization module 10 requests the PCF 30 to issue a new policy plan based on the policy optimization recommendation and updates the configuration parameters of the IP multimedia subsystem based on the new policy plan.

[0125] Specifically, based on the user behavior pattern, it is possible to predict the change in the user's service demand (such as when high bandwidth demand may occur), identify the user's long-term usage pattern (such as video calls or high bandwidth applications at a fixed time period every day), and then optimize the network resource allocation in advance (such as pre-adjusting the policy before the peak period to avoid the decline in network quality) and dynamically adjust the QoS policy (such as optimizing bandwidth allocation, priority control, coding adjustment, etc. for different user groups).

[0126] In a specific example, the user behavior pattern may include the key features in the following list:

[0127] Category Feature data Example Time feature Active time period of high-bandwidth services A user watches 4K videos every day from 19:00 to 22:00 Spatial feature Service access location A user often makes VoIP calls at the subway station Application feature Service type, usage frequency A user often uses high-definition video streaming media Device feature Device type, connection mode Mobile phone / tablet, WiFi / 5G Historical network status Network conditions at the same time and location in the past Insufficient bandwidth during a certain period in the past week

[0128] The NWDAF 40 can model these key features through LSTM (Long Short-Term Memory Network) or GRU (Gated Recurrent Unit) to predict the service demand and potential impact of a certain user / area in the future.

[0129] According to the predicted user behavior pattern, the intelligent analysis and prediction module 20 can perform proactive policy optimization. The following describes the main optimization methods in combination with the prediction examples in the above list:

[0130] (1) Optimization for high-bandwidth users

[0131] Prediction result: A certain user often watches 4K videos from 19:00 to 22:00, with high bandwidth requirements.

[0132] Optimization strategy: Reserve network resources in advance and perform adaptive bitrate control.

[0133] Reserving network resources in advance is specifically as follows:

[0134] Before 19:00, allocate a higher QoS policy for this user in advance to avoid congestion during peak hours from affecting the experience.

[0135] In areas where congestion may occur (such as residential areas and shopping malls), allocate higher bandwidth in advance.

[0136] Adaptive bitrate control is specifically as follows:

[0137] If it is predicted that network congestion may occur in this area in the future, it is recommended to reduce the video bitrate of this user (such as from 4K to 1080p).

[0138] (2) Optimization for real-time communication services (VoIP / video calls)

[0139] Prediction result: A certain user makes long VoIP calls in the company from 10:00 to 11:00 every morning.

[0140] Optimization strategy: Prioritize the allocation of low-latency paths and intelligent QoS dynamic adjustment.

[0141] The specific allocation of low-latency paths is as follows:

[0142] When it is predicted that there may be a call demand during this time period, adjust the routing in advance and select a low-latency link to ensure voice quality.

[0143] The specific intelligent QoS dynamic adjustment is as follows:

[0144] If it is predicted that the network condition may deteriorate during the call, adjust the QoS policy in advance (such as increasing the priority of VoIP data to avoid packet loss).

[0145] (3) Optimization for mobile users (such as high-speed rail / subway scenarios)

[0146] Prediction result: A certain user makes video calls on the subway from 08:00 to 09:00 every day, and the signal switches frequently.

[0147] Optimization strategy: Optimize the handover strategy and intelligent load balancing in advance.

[0148] The specific pre-optimization of the handover strategy is as follows:

[0149] Predict that the user will move from base station A to base station B, and optimize the resource allocation of base station B in advance to reduce packet loss during handover.

[0150] The specific intelligent load balancing is as follows:

[0151] If it is predicted that the load of multiple base stations on a certain line is too high, it is recommended to switch some users to WiFi hotspots or other frequency bands to reduce the pressure on the cellular network.

[0152] (4) Optimization for low-priority services (saving critical resources)

[0153] Prediction result: During peak hours in a certain area (such as a commercial center), there is a large amount of low-priority traffic (such as P2P downloads, background updates), which affects high-priority services.

[0154] Optimization strategy: Throttle or schedule low-priority services, specifically:

[0155] When it is predicted that the bandwidth is tight during peak hours, it is recommended to throttle low-priority services (such as P2P traffic, system updates).

[0156] Through intelligent scheduling, these services are postponed to off-peak hours (such as early morning).

[0157] In some embodiments, user behavior data can also be combined with network prediction data and / or congestion probability prediction data predicted by the intelligent analysis and prediction module 20 for use to obtain a decision result.

[0158] Exemplary:

[0159] NWDAF 40 predicts that congestion will occur in this area within the next 5 minutes, and a certain user is making a high-priority VoIP call.

[0160] Optimization strategy: Immediately increase the QoS priority of this user and reduce the bandwidth occupancy of other non-critical services.

[0161] NWDAF 40 predicts that a certain user will watch high-definition videos for a long time at night, but historical data shows that the network in this area may experience a load peak during this period.

[0162] Optimization strategy: Perform load balancing in advance at the core network or edge computing layer to reduce possible network bottlenecks.

[0163] To better understand this application, the following description is provided in combination with the following examples:

[0164] Strategy Request Interface SRI:

[0165] Example of input parameters (sent from the policy optimization module 10 to the PCF 30):

[0166]

[0167] Output parameters (returned from the PCF 30 to the policy optimization module 10):

[0168]

[0169] As can be seen from the above examples, the parameters sent by the policy request interface SRI to the PCF 30 may include: user ID, service type, initial bandwidth requirement, current network load, QoS requirements (including latency and packet loss rate). The parameters returned by the PCF 30 to the policy optimization module 10 through the policy request interface SRI may include: allocated bandwidth, user priority, encoding policy, and re-evaluation interval.

[0170] Policy update interface SUI:

[0171] When network prediction shows that congestion is imminent or the user's QoE drops, etc., the policy optimization module 10 requests a policy update from the PCF 30 through the policy update interface SUI:

[0172] Input parameters:

[0173]

[0174] Output parameters (PCF 30 issues a new policy):

[0175]

[0176] As can be seen from the above examples, the parameters sent by the policy update interface SUI to the PCF 30 may include: user ID, current QoE, predicted network status (such as reduced bandwidth), requirement type (such as increasing user priority). The parameters returned by the PCF 30 to the policy optimization module 10 through the policy update interface SUI may include: allocated bandwidth, user priority, encoding policy, and re-evaluation interval.

[0177] Policy feedback interface SFI:

[0178] After the session ends or after the policy has been executed for some time, the policy optimization module 10 feeds back the execution status and the final QoE measurement value to the PCF 30 for the PCF 30 to optimize future policies.

[0179] Input parameters:

[0180]

[0181]

[0182] Without the PCF 30 returning data, the policy rule library can be updated in the background.

[0183] As can be seen from the above examples, the parameters sent by the policy feedback interface SFI to the PCF 30 may include: user ID, final QoE, average latency, average packet loss rate.

[0184] Data acquisition interface DGI:

[0185] Input parameters: Periodically obtain network status data from NWDAF 40:

[0186]

[0187] As can be seen from the above example, the data obtained by the data acquisition interface DGI from NWDAF 40 may include: timestamp, region ID, bandwidth utilization rate, congestion prediction situation, average round-trip time, and packet loss rate trend.

[0188] Result output interface PRI (the interface for the intelligent analysis and prediction module 20 to output data to the policy optimization module 10):

[0189] Output parameters:

[0190]

[0191]

[0192] As can be seen from the above example, the data output by the result output interface PRI to the policy optimization module 10 may include: user ID, recommended operation (such as reducing the bit rate), QoE change prediction, and confidence level.

[0193] The following is described in combination with algorithm and pseudo-code examples, which is only for facilitating the understanding of this application and should not be used as a limitation to this application.

[0194] Algorithm name: DPDA (Dynamic Policy Decision Algorithm) Dynamic Policy Decision Algorithm

[0195] Algorithm objective: Based on the data provided by NWDAF 40, discover the trend of network status deterioration in advance and output policy update suggestions to the policy optimization module 10.

[0196] Input:

[0197] NWDAF data sequence: bandwidth utilization rate, delay, packet loss rate, congestion probability prediction data

[0198] User's current policy and user's current QoE: CurrentQoE, CurrentPolicy

[0199] User priority: UserPriority

[0200] Output:

[0201] Policy advice object: PolicyAdvice (recommend reducing the code rate, increasing the priority, maintaining the status quo, etc.)

[0202] Pseudo-code description:

[0203]

[0204]

[0205]

[0206] It can be understood that the above algorithms and pseudo - codes are only for exemplary illustration. The focus of this application is that: the IP multimedia subsystem is additionally provided with a policy optimization module and an intelligent analysis and prediction module, and the policy optimization method executed based on these two modules, rather than the specific manner of the algorithm. This application can adopt various different algorithms, and various different combinations can also be used for the predicted data, etc.

[0207] The above - disclosed are only the preferred examples of this application, and the scope of rights of this application cannot be limited thereby. Therefore, all equivalent changes made according to the claims of this application fall within the scope covered by this application.

Claims

1. An IP multimedia subsystem, characterized in that: Including strategy optimization module and intelligent analysis and prediction module, The strategy optimization module is configured as follows: Before the session is established, a policy request is sent to the PCF to request the PCF to provide initial policy parameters; During the session, request the PCF to issue a new policy solution; The intelligent analysis and prediction module is configured as follows: Get current and historical network status data from NWDAF; Obtaining network prediction data based on the network status data prediction; Based on the network prediction data, a decision is made on whether a policy update is needed and a policy update suggestion is output to the policy optimization module. The policy optimization module determines whether to request the PCF to issue a new policy solution based on the policy update suggestion and requests the PCF to issue a new policy solution based on the policy update suggestion when it is determined that the PCF should be requested to issue a new policy solution.

2. The IP multimedia subsystem according to claim 1, characterized in that: The policy optimization module includes a policy request interface, a policy update interface and a policy feedback interface; The policy request interface is configured to: before the session is established, send a policy request to the PCF to request the PCF to provide the initial policy parameters; The policy update interface is configured to: request the PCF to issue a new policy solution during a session; The policy feedback interface is configured to: after the policy execution is completed, feedback the policy execution status to the PCF; The intelligent analysis and prediction module includes a data acquisition interface and a decision result interface; The data acquisition interface is configured to acquire network status data from the NWDAF; The result output interface is configured to send the policy update suggestion to the policy optimization module.

3. The IP multimedia subsystem according to claim 1, characterized in that: The network status analysis data includes: bandwidth utilization, historical data flow, delay and packet loss rate; The network prediction data includes: bandwidth utilization prediction data, delay prediction data, and packet loss rate prediction data; The intelligent analysis and prediction module is further configured to: obtain congestion probability prediction data from the NWDAF; The intelligent analysis and prediction module makes a decision on whether a policy update is needed based on the network prediction data and the congestion probability prediction data and outputs the policy update suggestion.

4. The IP multimedia subsystem according to claim 3, characterized in that: The intelligent analysis and prediction module is further configured to: obtain the user's current strategy, the user's current QoE and the user's priority; The intelligent analysis and prediction module makes a decision on whether a policy update is needed and outputs the policy update suggestion based on the network prediction data, the congestion probability prediction data, the user's current policy, the user's current QoE and the user priority.

5. The IP multimedia subsystem according to claim 1, characterized in that: The IP multimedia subsystem includes a CSCF, an MRF and an AS, and the policy optimization module dynamically adjusts the session policy based on the signaling control of the CSCF and the media processing information of the MRF; The intelligent analysis and prediction module obtains business data from the AS and mines user behavior patterns and predicts changes in service requirements based on the business data.

6. A method for optimizing a session strategy of an IP multimedia subsystem according to claim 1 or 2, characterized in that: The steps include: S1, the IP multimedia subsystem receives a session establishment request from a UE; S2, the policy optimization module requests the PCF to provide the initial policy parameters and receives the initial policy parameters returned by the PCF to configure the IP multimedia subsystem to execute the initial policy; S3, the intelligent analysis and prediction module obtains the network status data from the NWDAF and obtains the network prediction data based on the network status data; S4, the intelligent analysis and prediction module makes a decision on whether a policy update is needed based on the network prediction data and outputs the policy update suggestion to the policy optimization module; S5, the policy optimization module determines whether to request the PCF to issue a new policy solution according to the policy update suggestion, and if the determination result is yes, proceeds to step S6; S6, the policy optimization module requests the PCF to issue a new policy solution based on the policy update suggestion and updates the configuration parameters of the IP multimedia subsystem based on the new policy solution; S7: The IP multimedia subsystem executes the new policy.

7. The method according to claim 6, characterized in that After step S7, the method further includes: S8, after the current session ends, the policy optimization module sends the policy execution status to the PCF; In step S5, if the determination result is no, then go to step S8.

8. The method according to claim 6, characterized in that The intelligent analysis and prediction module makes a decision on whether a policy update is needed based on the network prediction data and outputs the policy update suggestion to the policy optimization module. The intelligent analysis and prediction module includes: The network prediction data is compared with the prediction data threshold, and a decision is made on whether a policy update is needed based on the comparison result, and the policy update suggestion is output to the policy optimization module.

9. The method according to claim 6, characterized in that The input parameters of the intelligent analysis and prediction module also include: NWDAF prediction data, user current strategy, user current QoE and user priority; The intelligent analysis and prediction module makes a decision on whether a policy update is needed based on the network prediction data, the NWDAF prediction data, the user's current policy, the user's current QoE and the user priority, and outputs the policy update suggestion to the policy optimization module.

10. The method according to claim 6, characterized in that Also includes: The NWDAF predicts the user behavior pattern based on the historical data and real-time data of the user behavior pattern to obtain the predicted data of the user behavior pattern; The intelligent analysis and prediction module makes a decision on whether policy optimization is needed based on the prediction data of the user behavior pattern obtained from the NWDAF and outputs the policy optimization suggestion to the policy optimization module; The policy optimization module requests the PCF to issue a new policy solution based on the policy optimization suggestion and updates the configuration parameters of the IP multimedia subsystem based on the new policy solution.

Images