A method and device for data analysis based on NWDAF and a readable storage medium

Abstract

The application provides a kind of NWDAF-based data analysis method, device and readable storage medium, the method comprises: receiving the customized data analysis request sent by application function AF;According to AF ID and ML Model ID, the corresponding target ML Model is obtained;According to the data input information required for customized data analysis, data analysis is carried out using the target ML Model, and the data analysis result corresponding to the analysis output information expected by AF is obtained;The customized data analysis response message carrying the data analysis result is returned to AF.This method, device and readable storage medium can solve the problem that the existing standardized data analysis method is limited to the defined 15 analysis types / analysis scenarios, resulting in insufficient flexibility of data analysis, and difficulty in meeting the user demand under the condition of new business / new scene.

Description

Technical Field

[0001] This invention relates to the field of network technology, and in particular to a data analysis method, apparatus and readable storage medium based on NWDAF. Background Technology

[0002] 3GPP TS 23.288 defines 15 analysis types / scenarios, including UE Communication Analysis (Analytics ID: UE Communication), Abnormal Behavior Analysis (Analytics ID: Abnormal Behaviour), and User Data Congestion Analysis (Analytics ID: User Data Congestion), as well as the data inputs and outputs required for each analysis type / scenarios, providing standardized intelligent analysis solutions for these scenarios. However, this high degree of standardization also brings a problem: the data analysis based on the 5G core network lacks flexibility, and analysis behavior is limited to the 15 defined analysis types / scenarios. Furthermore, 3GPP's standardization cycle is relatively long (generally 1.5 to 2 years for a release), and with the rapid development of mobile communication services, standardized data analysis solutions are insufficient to meet the ever-changing user needs of new services and scenarios. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to address the above-mentioned shortcomings of the prior art by providing a data analysis method, device and readable storage medium based on NWDAF, so as to solve the problem that the existing standardized data analysis methods are limited to 15 predefined analysis types / scenarios, resulting in insufficient flexibility of data analysis and difficulty in meeting the user needs of new businesses / new scenarios that are changing rapidly.

[0004] In a first aspect, the present invention provides a data analysis method based on NWDAF, applied to a target NWDAF, the method comprising:

[0005] The application function (AF) sends a customized data analysis request. The customized data analysis request carries the analysis identifier (Analytics ID), the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured in the machine learning model (ML Model) required for the customized data analysis, the ML Model ID of the ML Model, the data input information required for the customized data analysis, and the analysis output information expected by the AF. The value of the Analytics ID is a custom value.

[0006] Obtain the corresponding target ML Model based on the AF ID and ML Model ID;

[0007] Based on the data input information required for the customized data analysis, the target ML Model is used to perform data analysis to obtain the data analysis results corresponding to the analysis output information expected by the AF.

[0008] Return a customized data analysis response message carrying the data analysis results to the AF.

[0009] Furthermore, the data input information required for the customized data analysis includes the data types to be collected and the corresponding data source types.

[0010] Further, obtaining the corresponding target ML Model based on the AF ID and ML Model ID specifically includes:

[0011] Determine whether the target NWDAF has the target MLModel pre-stored based on the AF ID and ML Model ID;

[0012] If so, then retrieve the pre-stored target ML Model;

[0013] If not, an ML Model retrieval request is sent to the NWDAF corresponding to the NWDAF ID, so that the NWDAF corresponding to the NWDAF ID returns the target ML Model to the target NWDAF based on the ML Model retrieval request.

[0014] Furthermore, there are multiple NWDAF IDs, and sending an ML Model retrieval request to the NWDAF corresponding to the NWDAF ID specifically includes:

[0015] Select one NWDAF from multiple NWDAF IDs according to a pre-configured local policy;

[0016] Send an ML Model retrieval request to the selected NWDAF.

[0017] Furthermore, before obtaining the pre-stored target ML Model, the method further includes:

[0018] The system receives a customized ML Model configuration request sent by the AF through the Network Open Function (NEF). The customized ML Model configuration request carries the AF ID, the target ML Model that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target ML Model, ML Model filtering information, and ML Model reporting target.

[0019] According to the customized ML Model configuration request, store the target ML Model and the corresponding AF ID, ML Model ID, ML Model filtering information and ML Model reporting target of the target ML Model;

[0020] The NEF returns a customized ML Model configuration response message to the AF, and the customized ML Model configuration response message carries the NWDAF ID of the target NWDAF.

[0021] Furthermore, the step of performing data analysis using the target ML Model based on the data input information required for the customized data analysis to obtain the data analysis results corresponding to the analysis output information expected by the AF specifically includes:

[0022] Data corresponding to the data type is collected from the data source corresponding to the data source type, and the collected data is analyzed using the target ML Model to obtain the data analysis results corresponding to the analysis output information expected by the AF.

[0023] Secondly, this invention provides a data analysis method based on NWDAF, applied to application function AF, the method comprising:

[0024] Send a customized data analysis request to the target NWDAF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured by the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value.

[0025] The system receives a customized data analysis response message returned by the target NWDAF, which carries the data analysis results. The customized data analysis response message is sent by the target NWDAF after obtaining the corresponding target ML Model based on the AF ID and ML Model ID, performing data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtaining the data analysis results corresponding to the analysis output information expected by the AF.

[0026] Furthermore, the data input information required for the customized data analysis includes the data types to be collected and the corresponding data source types.

[0027] Furthermore, before sending the customized data analysis request to the target NWDAF, the method further includes:

[0028] The Network Open Function (NEF) sends a customized ML Model configuration request to the target NWDAF. The customized ML Model configuration request carries the AF ID, the target MLModel that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target MLModel, ML Model filtering information, and ML Model reporting target.

[0029] Receive the customized ML Model configuration response message returned by the target NWDAF through the NEF, the customized ML Model configuration response message carrying the NWDAF ID of the target NWDAF.

[0030] Furthermore, the step of sending a customized ML Model configuration request to the target NWDAF via the Network Open Function (NEF) specifically includes:

[0031] A customized ML Model configuration request is sent to the NEF so that the NEF requests the Network Repository Function (NRF) to discover an NWDAF with Model Training Logic Function (MTLF), and after receiving the target NWDAF returned by the NRF that matches the ML Model filtering information, the NEF sends the customized ML Model configuration request to the target NWDAF.

[0032] Thirdly, the present invention provides a data analysis device based on NWDAF, disposed at a target NWDAF, the device comprising:

[0033] The analysis request receiving module is used to receive customized data analysis requests sent by the application function AF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured in the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value.

[0034] The target model acquisition module, connected to the analysis request receiving module, is used to acquire the corresponding target ML Model based on the AF ID and MLModel ID;

[0035] The data analysis module, connected to the target model acquisition module, is used to perform data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtain the data analysis results corresponding to the analysis output information expected by the AF.

[0036] The analysis result sending module, connected to the data analysis module, is used to return a customized data analysis response message carrying the data analysis results to the AF.

[0037] Fourthly, the present invention provides a data analysis device based on NWDAF, configured in the application function AF, the device comprising:

[0038] The analysis request sending module is used to send a customized data analysis request to the target NWDAF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured by the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value.

[0039] The analysis result receiving module, connected to the analysis request sending module, is used to receive a customized data analysis response message returned by the target NWDAF, which carries the data analysis results. The customized data analysis response message is sent by the target NWDAF after obtaining the corresponding target ML Model based on the AF ID and ML Model ID, performing data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtaining the data analysis results corresponding to the analysis output information expected by the AF.

[0040] Fifthly, the present invention provides a data analysis device based on NWDAF, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement the NWDAF-based data analysis method described in the first or second aspect above.

[0041] In a sixth aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the NWDAF-based data analysis method described in the first or second aspect above.

[0042] This invention provides a data analysis method, apparatus, and readable storage medium based on NWDAF. First, it receives a customized data analysis request sent by an Application Function (AF). This request carries an Analysis Identifier (Analytics ID), the AF ID corresponding to the AF, the NWDAFID of the NWDAF configured for the Machine Learning Model (ML Model) required for the customized data analysis, the ML Model ID of the ML Model, the data input information required for the customized data analysis, and the analysis output information expected by the AF. The Analysis ID is a custom value. Then, it obtains the corresponding target ML Model based on the AF ID and ML ModelID. Next, it performs data analysis using the target ML Model based on the data input information required for the customized data analysis, obtaining the data analysis result corresponding to the analysis output information expected by the AF. Finally, it returns a customized data analysis response message carrying the data analysis result to the AF. This invention, through customized data analysis using NWDAF, improves the flexibility of data analysis in the 5G core network. It solves the problem that existing standardized data analysis methods, which are limited to 15 predefined analysis types / scenarios, lack flexibility and struggle to meet the rapidly evolving user needs of new services and scenarios. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the data analysis architecture based on NWDAF;

[0044] Figure 2 A flowchart illustrating a typical NWDAF service consumer requesting analysis information from NWDAF;

[0045] Figure 3 This is a flowchart of a data analysis method based on NWDAF according to Embodiment 1 of the present invention;

[0046] Figure 4 This is an interactive diagram illustrating a customized ML Model delivery process based on NWDAF according to an embodiment of the present invention.

[0047] Figure 5 This is an interactive schematic diagram of a data analysis method based on NWDAF according to an embodiment of the present invention;

[0048] Figure 6 This is a flowchart of a data analysis method based on NWDAF according to Embodiment 2 of the present invention;

[0049] Figure 7 This is a schematic diagram of the structure of a data analysis device based on NWDAF according to Embodiment 3 of the present invention;

[0050] Figure 8 This is a schematic diagram of the structure of a data analysis device based on NWDAF according to Embodiment 4 of the present invention;

[0051] Figure 9 This is a schematic diagram of the structure of a data analysis device based on NWDAF according to Embodiment 5 of the present invention. Detailed Implementation

[0052] To enable those skilled in the art to better understand the technical solution of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0053] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0054] It is understood that, without conflict, the various embodiments and features in the embodiments of the present invention can be combined with each other.

[0055] It is understood that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, while the parts unrelated to the present invention are not shown in the drawings.

[0056] It is understood that each unit or module involved in the embodiments of the present invention may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple units or modules may be integrated into one entity structure.

[0057] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this invention may occur in a different order than that marked in the accompanying drawings.

[0058] It is understood that the flowcharts and block diagrams of this invention illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this invention. Each block in the flowchart or block diagram may represent a unit, module, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagram and flowchart can be implemented using a hardware-based system to achieve the specified function, or using a combination of hardware and computer instructions.

[0059] It is understood that the units and modules involved in the embodiments of the present invention can be implemented by software or by hardware. For example, the units and modules can be located in a processor.

[0060] Application Overview

[0061] The 3rd Generation Partnership Project (3GPP), an international telecommunications standards organization, has defined a 5G-native big data analytics and AI technology application framework based on the Network Data Analytics Function (NWDAF) within the framework of the 5G core network (5GC). The NWDAF architecture supports standardized communication network processes for data collection, data analysis, and feedback of analysis results, further enhancing the automated management, operation, and optimization capabilities of 5G networks.

[0062] Figure 1 A schematic diagram of the NWDAF-based data analysis architecture is shown, as follows: Figure 1 As shown, NWDAF can obtain mobility management information such as user location from core network control plane network functions (NFs), such as Access and Mobility Management Function (AMF), and session management information such as the type of service currently or previously used by the user from Session Management Function (SMF). It can also obtain NF registration information from Network Repository Function (NRF), and also supports obtaining information such as network resource usage or NF load from the network management OAM.

[0063] A typical process of a NWDAF service consumer requesting analytics information from NWDAF is as follows: Figure 2 As shown, it includes the following steps:

[0064] 1) The NWDAF service consumer sends an analysis information request message (Nnwdaf_AnalyticsInfo_Request) to NWDAF. The message contains the type of analysis information that the NWDAF service consumer wants to obtain (marked by the analysis identifier, i.e., Analytics ID), as well as other limiting conditions for the analysis information, such as whether the analysis is based on statistics of past data or prediction of future situations, the time window corresponding to the analysis information, etc.

[0065] 2) After completing the analysis, NWDAF sends a response message (Nnwdaf_AnalyticsInfo_Request response) to the NWDAF service consumer in response to the analysis information request. The message contains the analysis results information.

[0066] In this context, NWDAF service consumers can be other 5G core network functions besides NWDAF, such as AMF, SMF, PCF, and AF (Application Function). 3GPP TS 23.288 defines several analysis types / scenarios and corresponding Analytics IDs. Based on the Analytics ID requested by the NWDAF service consumer, NWDAF determines the requested analysis type / scenarios and collects data from different data source NFs. Simultaneously, it applies different ML Models (Machine Learning Models) to perform data analysis to meet the NWDAF service consumer's request. 3GPP TS 23.288 defines standardized data input and analysis output templates for each Analytics ID. Upon receiving the Analytics ID carried in the NWDAF service consumer's request message, NWDAF can clearly identify which data sources to collect, which data to collect, and which data analysis results to ultimately output.

[0067] However, while 3GPP TS 23.288 defines 15 analysis types / scenarios, such as UE Communication (Analytics ID: UE Communication), Abnormal Behavior (Analytics ID: Abnormal behaviour), and User Data Congestion (Analytics ID: User Data Congestion), and provides standardized intelligent analysis solutions for these scenarios, this high degree of standardization also brings a problem: insufficient flexibility in data analysis based on the 5G core network, limiting analysis behavior to the defined 15 analysis types / scenarios. Furthermore, 3GPP's standardization cycle is relatively long (typically 1.5 to 2 years for a single release), and with the rapid development of mobile communication services, standardized data analysis solutions struggle to meet the ever-evolving user needs of new services and scenarios.

[0068] To address the aforementioned technical issues, this application proposes a data analysis method, apparatus, and readable storage medium based on NWDAF. It defines a mechanism for AF to request customized data analysis. AF can issue customized data analysis requests to the target NWDAF based on its own business scenarios and actual business needs. These requests include the data input information required for customized data analysis and the analysis output information expected by the AF, enabling the target NWDAF to perform customized data analysis and obtain the data analysis results corresponding to the analysis output information expected by the AF. This improves the flexibility of data analysis in the 5G core network.

[0069] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0070] Example 1:

[0071] This embodiment provides a data analysis method based on NWDAF, applied to the target NWDAF, such as... Figure 3 As shown, the method includes:

[0072] Step S101: Receive a customized data analysis request sent by the application function AF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured in the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value.

[0073] In this embodiment, the target NWDAF has AnLF (Analytics logical function) functionality. The NWDAF ID can be one or more, and the custom value can be any value other than the Analytics ID defined in 3GPP TS 23.288, such as "Customized".

[0074] Optionally, the data input information required for the customized data analysis includes the data types and corresponding data source types to be collected for the customized data analysis. The data types and corresponding data source types can be shown in Table 1. The expected analysis output information of AF indicates the analysis results that AF expects NWDAF to obtain based on the collected data. The analysis results can be shown in Table 2. It should be noted that the input and output information in Tables 1 and 2 are only examples; in actual applications, AF can be customized based on analysis requirements.

[0075] Table 1: Data Types and Corresponding Data Source Types

[0076]

[0077]

[0078] Table 2: Analysis Results

[0079]

[0080] Step S102: Obtain the corresponding target ML Model based on the AF ID and ML Model ID.

[0081] Specifically, the target NWDAF determines whether it has pre-stored the target ML Model based on the AF ID and ML Model ID; if so, it retrieves the pre-stored target ML Model; if not, it sends an ML Model retrieval request to the NWDAF corresponding to the NWDAF ID, so that the NWDAF corresponding to the NWDAF ID returns the target ML Model to the target NWDAF based on the ML Model retrieval request.

[0082] In this embodiment, the target NWDAF internally retrieves the acquired ML Model information based on the combination of AF ID and ML Model ID in the customized data analysis request. If the target NWDAF already contains the corresponding target ML Model (e.g., the AF has previously requested the same customized analysis), then the target ML Model is acquired. If the target NWDAF does not contain the corresponding target ML Model, then the NWDAF (MTLF) storing the required target ML Model is discovered based on the NWDAF ID received in the customized data analysis request. This NWDAF (MTLF) discovery process may be completed through an NRF network element.

[0083] Optionally, there are multiple NWDAF IDs, and sending an ML Model retrieval request to the NWDAF corresponding to the NWDAF ID specifically includes:

[0084] Select one NWDAF from multiple NWDAF IDs according to a pre-configured local policy;

[0085] Send an ML Model retrieval request to the selected NWDAF.

[0086] In this embodiment, the AF can send multiple NWDAF IDs. If there are multiple NWDAF IDs, the target NWDAF selects an NWDAF (MTLF) according to its local configuration policy, such as selecting an NWDAF (MTLF) with a larger management area, and sends an ML Model retrieval request to the selected NWDAF (MTLF). The request includes the AF ID and the ML Model ID. The NWDAF (MTLF) retrieves the ML Model based on the ML Model retrieval request received from the target NWDAF and returns an ML Model retrieval response message to the target NWDAF. This message includes the target ML Model requested by the target NWDAF, the ML Model filtering information corresponding to the target ML Model (indicating the scope to which the ML Model applies, such as the slice identifier S-NSSAI of a specific slice, and the location area information of the ML Model application), and the ML Model reporting target (indicating the object of data analysis, such as a specific UE, a group of UEs, or any UE).

[0087] Optionally, before obtaining the pre-stored target ML Model, the method further includes:

[0088] The system receives a customized ML Model configuration request sent by the AF through the Network Open Function (NEF). The customized ML Model configuration request carries the AF ID, the target ML Model that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target ML Model, ML Model filtering information, and ML Model reporting target.

[0089] According to the customized ML Model configuration request, store the target ML Model and the corresponding AF ID, ML Model ID, ML Model filtering information and ML Model reporting target of the target ML Model;

[0090] The NEF returns a customized ML Model configuration response message to the AF, and the customized ML Model configuration response message carries the NWDAF ID of the target NWDAF.

[0091] In this embodiment, the AF sends a customized ML Model configuration request to the NEF, so that the NEF requests the Network Repository Function (NRF) to discover an NWDAF with Model Training Logic Function (MTLF), and after receiving the target NWDAF returned by the NRF that matches the ML Model filtering information, the NEF sends the customized ML Model configuration request to the target NWDAF.

[0092] Specifically, refer to Figure 4 This diagram illustrates an interactive schematic of a customized ML Model delivery process based on NWDAF, provided by an embodiment of the present invention. This embodiment includes the following steps:

[0093] (1) The AF sends a customized ML Model configuration request to the NEF (Network Exposure Function). The request contains the AF ID corresponding to the AF, the ML Model that the AF wants to configure in the NWDAF (MTLF), the ML Model ID, the ML Model filtering information (indicating the scope of the ML Model, such as the slice identifier S-NSSAI of a specific slice, the location area information to which the ML Model applies), and the ML Model reporting target (indicating the object of data analysis, such as a specific UE, a group of UEs, or any UE). Among these, the ML Model reporting target is an optional parameter that needs to be carried when the analysis type is related to a specific UE, while the others are mandatory parameters.

[0094] (2) The NEF sends an NWDAF (MTLF) discovery request to the NRF to request the discovery of an NWDAF with MTLF (Model Training logical function) functionality. The request contains ML Model filtering information sent by the AF.

[0095] (3) The NRF selects the appropriate NWDAF(MTLF) based on the ML Model filtering information contained in the NWDAF(MTLF) discovery request sent by the NEF, and returns the NWDAF(MTLF) that conforms to the ML Model filtering information sent by the NEF to the NEF through the NWDAF(MTLF) discovery response message.

[0096] It should be noted that NRF can return multiple NWDAFs (MTLFs) to NEF. For example, if the analysis area covered by a single NWDAF (MTLF) is smaller than the location area of ​​the ML Model application specified in the ML Model filter information, multiple NWDAFs (MTLFs) are needed to cover the complete location area of ​​the ML Model application specified in the ML Model filter information.

[0097] (4) The NEF sends a customized ML Model configuration request to the NWDAF (MTLF). This request includes the ML Model to be configured in the NWDAF (MTLF), the AF ID of the AF that issued the ML Model, the ML Model ID, the ML Model filtering information (indicating the scope to which the ML Model applies, such as the slice identifier S-NSSAI for a specific slice, and the location area information for which the ML Model is applied), and the ML Model reporting target (indicating the object of data analysis, such as a specific UE, a group of UEs, or any UE). If the NRF returns multiple NWDAFs (MTLFs) to the NEF, then the NEF sends a customized ML Model configuration request containing the above information to each NWDAF (MTLF) returned by the NRF.

[0098] (5) NWDAF(MTLF) returns a customized ML Model configuration response message to NEF, indicating that the customized ML Model configuration is successful. The response message carries the NWDAF ID of each NWDAF(MTLF).

[0099] (6) NEF returns a customized ML Model configuration response message to AF, indicating that the customized ML Model configuration is successful. The message carries the NWDAF ID of each NWDAF (MTLF).

[0100] Step S103: Based on the data input information required for the customized data analysis, perform data analysis using the target MLModel to obtain the data analysis results corresponding to the analysis output information expected by the AF.

[0101] Specifically, the target NWDAF collects data corresponding to the data type from the data source corresponding to the data source type, and uses the target ML Model to perform data analysis on the collected data to obtain the data analysis results corresponding to the analysis output information expected by the AF.

[0102] In this embodiment, the data acquisition, analysis, and result feedback process can be the standard process defined in 3GPP TS 23.288.

[0103] Step S104: Return a customized data analysis response message carrying the data analysis results to the AF.

[0104] In one specific embodiment, reference Figure 5 This diagram illustrates an interactive schematic of a data analysis method based on NWDAF provided by an embodiment of the present invention. NWDAF (MTLF) and NWDAF (AnLF) may or may not be combined. In this embodiment, the following steps are included:

[0105] (1) The AF sends a customized data analysis request to the NWDAF with AnLF (Analytics logical function). The request includes the Analytics ID (in the case of customized analysis, the Analytics ID is "Customized"). The AF also carries the AF ID corresponding to the AF, the NWDAFID of the NWDAF (MTLF) configured in the ML Model of the requested customized data analysis, the ML Model ID of the ML Model, the data input required for customized analysis and the analysis output expected by the AF in the customized data analysis request message. The data input required for analysis indicates the data type and the corresponding data source type that the customized data analysis needs to collect.

[0106] (2) NWDAF(AnLF) retrieves the obtained ML Model information internally based on the combination of AF ID + ML Model ID in the customized data analysis request. If there is already a corresponding ML Model in NWDAF(AnLF) (such as AF having requested the same customized analysis before), then skip steps (3) and (4); otherwise, execute steps (3) and (4).

[0107] (3) The NWDAF(AnLF) discovers the NWDAF(MTLF) storing the required ML Model based on the NWDAF ID received in the customized data analysis request in step (1). This NWDAF(MTLF) discovery process may be completed through the NRF network element. If multiple NWDAF IDs are sent in step (1), the NWDAF(AnLF) selects one NWDAF(MTLF) according to its local configuration policy. The NWDAF(AnLF) sends an ML Model retrieval request to the selected NWDAF(MTLF), which includes the AF ID and the ML Model ID.

[0108] (4) The NWDAF (MTLF) retrieves the ML Model based on the ML Model retrieval request received from the NWDAF (AnLF) and returns an ML Model retrieval response message to the NWDAF (AnLF). This message contains the ML Model requested by the NWDAF (AnLF), ML Model filtering information (indicating the scope to which the ML Model applies, such as the slice identifier S-NSSAI of a specific slice, and the location area information to which the ML Model is applied), and ML Model reporting target (indicating the object of data analysis, such as a specific UE, a group of UEs, or any UE).

[0109] (5) NWDAF (AnLF) collects data from the corresponding data source based on the data type and corresponding data source type required for customized data analysis received from AF, and performs data analysis based on the ML Model obtained from NWDAF (MTLF), and finally forms the data analysis results requested by AF. The process of data collection, analysis and result feedback can be the standard process defined by 3GPP TS 23.288.

[0110] (6) NWDAF(AnLF) returns a customized data analysis response message to AF, which contains the data analysis results requested by AF.

[0111] The data analysis method based on NWDAF provided in this invention first receives a customized data analysis request sent by an application function (AF). This request carries an analysis identifier (Analytics ID), the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured for the machine learning model (ML Model) required for the customized data analysis, the ML Model ID of the ML Model, the data input information required for the customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value. Then, the corresponding target ML Model is obtained based on the AF ID and ML Model ID. Next, based on the data input information required for the customized data analysis, data analysis is performed using the target ML Model to obtain the data analysis result corresponding to the analysis output information expected by the AF. Finally, a customized data analysis response message carrying the data analysis result is returned to the AF. This invention, through customized data analysis using NWDAF, can improve the flexibility of data analysis in the 5G core network. It solves the problem that existing standardized data analysis methods, which are limited to 15 predefined analysis types / scenarios, lack flexibility and struggle to meet the rapidly evolving user needs of new services and scenarios.

[0112] Example 2:

[0113] like Figure 6 As shown, this embodiment provides a data analysis method based on NWDAF, applied to application function AF, the method including:

[0114] Step S201: Send a customized data analysis request to the target NWDAF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured in the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value.

[0115] In this embodiment, the target NWDAF has AnLF (Analytics logical function) functionality. The NWDAF ID can be one or more, and the custom value can be any value other than the Analytics ID defined in 3GPP TS 23.288, such as "Customized".

[0116] Optionally, the data input information required for the customized data analysis includes the data types to be collected and the corresponding data source types.

[0117] Optionally, before sending the customized data analysis request to the target NWDAF, the method further includes:

[0118] The Network Open Function (NEF) sends a customized ML Model configuration request to the target NWDAF. The customized ML Model configuration request carries the AF ID, the target MLModel that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target MLModel, ML Model filtering information, and ML Model reporting target.

[0119] Receive the customized ML Model configuration response message returned by the target NWDAF through the NEF, the customized ML Model configuration response message carrying the NWDAF ID of the target NWDAF.

[0120] Optionally, sending a customized ML Model configuration request to the target NWDAF via the Network Open Function (NEF) specifically includes:

[0121] A customized ML Model configuration request is sent to the NEF so that the NEF requests the Network Repository Function (NRF) to discover an NWDAF with Model Training Logic Function (MTLF), and after receiving the target NWDAF returned by the NRF that matches the ML Model filtering information, the NEF sends the customized ML Model configuration request to the target NWDAF.

[0122] Step S202: Receive a customized data analysis response message returned by the target NWDAF, which carries the data analysis results. The customized data analysis response message is sent by the target NWDAF after obtaining the corresponding target ML Model based on the AF ID and ML ModelID, performing data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtaining the data analysis results corresponding to the analysis output information expected by the AF.

[0123] Example 3:

[0124] like Figure 7 As shown, this embodiment provides a data analysis device based on NWDAF, which is set at the target NWDAF and used to execute the NWDAF-based data analysis method in Embodiment 1 above. The device includes:

[0125] Analysis request receiving module 11 is used to receive customized data analysis requests sent by application function AF. The customized data analysis request carries analysis identifier AnalyticsID, AF ID corresponding to the AF, NWDAF ID of NWDAF configured in the machine learning model ML Model required for customized data analysis, ML Model ID of ML Model, data input information required for customized data analysis, and analysis output information expected by AF. The value of AnalyticsID is a custom value.

[0126] The target model acquisition module 12 is connected to the analysis request receiving module 11 and is used to acquire the corresponding target ML Model based on the AF ID and MLModel ID.

[0127] Data analysis module 13, connected to target model acquisition module 12, is used to perform data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtain the data analysis results corresponding to the analysis output information expected by the AF.

[0128] The analysis result sending module 14 is connected to the data analysis module 13 and is used to return a customized data analysis response message carrying the data analysis results to the AF.

[0129] Optionally, the data input information required for the customized data analysis includes the data types to be collected and the corresponding data source types.

[0130] Optionally, the target model acquisition module 12 specifically includes:

[0131] The determination unit is used to determine whether the target NWDAF has the target ML Model pre-stored based on the AF ID and ML Model ID;

[0132] The first processing unit is configured to, if so, obtain the pre-stored target ML Model;

[0133] The second processing unit is configured to, if not, send an ML Model retrieval request to the NWDAF corresponding to the NWDAF ID, so that the NWDAF corresponding to the NWDAF ID returns the target ML Model to the target NWDAF based on the ML Model retrieval request.

[0134] Optionally, there are multiple NWDAF IDs, and the second processing unit specifically includes:

[0135] The selection unit is used to select one NWDAF from multiple NWDAF IDs according to a pre-set local configuration policy.

[0136] The retrieval request sending unit is used to send an ML Model retrieval request to the selected NWDAF.

[0137] Optionally, the device further includes:

[0138] The configuration request receiving module is used to receive the customized MLModel configuration request sent by the AF through the Network Open Function (NEF). The customized MLModel configuration request carries the AF ID, the target MLModel that the AF wants to configure in the target NWDAF, the MLModel ID corresponding to the target MLModel, MLModel filtering information, and MLModel reporting target.

[0139] The storage module is used to store the target ML Model, the AF ID, ML Model ID, ML Model filtering information, and ML Model reporting target corresponding to the target ML Model, according to the customized ML Model configuration request.

[0140] A configuration response sending module is configured to return a customized ML Model configuration response message to the AF through the NEF, wherein the customized ML Model configuration response message carries the NWDAF ID of the target NWDAF.

[0141] Optionally, the data analysis module 13 is specifically used for:

[0142] Data corresponding to the data type is collected from the data source corresponding to the data source type, and the collected data is analyzed using the target ML Model to obtain the data analysis results corresponding to the analysis output information expected by the AF.

[0143] Example 4:

[0144] like Figure 8 As shown, this embodiment provides a data analysis device based on NWDAF, configured in the application function AF, for executing the NWDAF-based data analysis method in embodiment 2 above. The device includes:

[0145] The analysis request sending module 21 is used to send a customized data analysis request to the target NWDAF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured by the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, and the data input information required for customized data analysis and the analysis output information expected by the AF. The Analytics ID is a custom value.

[0146] The analysis result receiving module 22, connected to the analysis request sending module 21, is used to receive a customized data analysis response message returned by the target NWDAF, which carries the data analysis results. The customized data analysis response message is sent by the target NWDAF after obtaining the corresponding target ML Model based on the AF ID and ML Model ID, performing data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtaining the data analysis results corresponding to the analysis output information expected by the AF.

[0147] Optionally, the data input information required for the customized data analysis includes the data types to be collected and the corresponding data source types.

[0148] Optionally, the device further includes:

[0149] The configuration request sending module is used to send a customized MLModel configuration request to the target NWDAF through the Network Open Function (NEF). The customized MLModel configuration request carries the AF ID, the target ML Model that the AF wants to configure in the target NWDAF, the ML ModelID corresponding to the target ML Model, ML Model filtering information, and ML Model reporting target.

[0150] A configuration response receiving module is configured to receive a customized ML Model configuration response message returned by the target NWDAF through the NEF, wherein the customized ML Model configuration response message carries the NWDAF ID of the target NWDAF.

[0151] Optionally, the configuration request sending module is specifically used for:

[0152] A customized ML Model configuration request is sent to the NEF so that the NEF requests the Network Repository Function (NRF) to discover an NWDAF with Model Training Logic Function (MTLF), and after receiving the target NWDAF returned by the NRF that matches the ML Model filtering information, the NEF sends the customized ML Model configuration request to the target NWDAF.

[0153] Example 5:

[0154] refer to Figure 9 This embodiment provides a data analysis device based on NWDAF, including a memory 31 and a processor 32. The memory 31 stores a computer program, and the processor 32 is configured to run the computer program to execute the NWDAF-based data analysis method in Embodiment 1 or Embodiment 2.

[0155] The memory 31 is connected to the processor 32. The memory 31 can be a flash memory, a read-only memory or other memory, and the processor 32 can be a central processing unit or a microcontroller.

[0156] Example 6:

[0157] This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the NWDAF-based data analysis method in Embodiment 1 or Embodiment 2 described above.

[0158] The computer-readable storage medium includes volatile or non-volatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program modules, or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer.

[0159] In summary, the data analysis method, apparatus, and readable storage medium based on NWDAF provided in this embodiment of the invention first receive a customized data analysis request sent by an application function (AF). The customized data analysis request carries an analysis identifier (Analytics ID), the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured in the machine learning model (ML Model) required for the customized data analysis, the ML Model ID of the ML Model, the data input information required for the customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value. Then, the corresponding target ML Model is obtained based on the AF ID and ML Model ID. Next, based on the data input information required for the customized data analysis, data analysis is performed using the target ML Model to obtain the data analysis result corresponding to the analysis output information expected by the AF. Finally, a customized data analysis response message carrying the data analysis result is returned to the AF. This invention improves the flexibility of data analysis in 5G core networks through customized data analysis using NWDAF. It solves the problem that existing standardized data analysis methods are limited to 15 predefined analysis types / scenarios, resulting in insufficient flexibility and difficulty in meeting the ever-changing user needs of new services / scenarios.

[0160] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A data analysis method based on Network Data Analysis Function (NWDAF), applied to target NWDAF, characterized in that, The method includes: The application function (AF) receives a customized machine learning model (ML Model) configuration request sent through the network open function (NEF). The customized ML Model configuration request carries the AF ID, the target ML Model that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target ML Model, ML Model filtering information, and ML Model reporting target. The ML Model filtering information is used to indicate the scope of application of the target ML Model, and the ML Model reporting target is used to indicate the object of data analysis. According to the customized ML Model configuration request, store the target ML Model and the corresponding AF ID, ML Model ID, ML Model filtering information and ML Model reporting target of the target ML Model; The NEF returns a customized ML Model configuration response message to the AF, and the customized ML Model configuration response message carries the NWDAF ID of the target NWDAF; The system receives a customized data analysis request sent by the AF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAFID of the NWDAF configured in the ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value. Obtain the corresponding target ML Model based on the AF ID and ML Model ID; Based on the data input information required for the customized data analysis, the target ML Model is used to perform data analysis to obtain the data analysis results corresponding to the analysis output information expected by the AF. Return a customized data analysis response message carrying the data analysis results to the AF.

2. The method according to claim 1, characterized in that, The data input information required for customized data analysis includes the data types to be collected and the corresponding data source types.

3. The method according to claim 1, characterized in that, The step of obtaining the corresponding target ML Model based on the AF ID and ML Model ID specifically includes: Determine whether the target NWDAF has the target ML Model pre-stored based on the AF ID and ML Model ID; If so, then retrieve the pre-stored target ML Model; If not, an ML Model retrieval request is sent to the NWDAF corresponding to the NWDAF ID, so that the NWDAF corresponding to the NWDAF ID returns the target ML Model to the target NWDAF based on the ML Model retrieval request.

4. The method according to claim 3, characterized in that, There are multiple NWDAF IDs, and sending an ML Model retrieval request to the NWDAF corresponding to the NWDAF ID specifically includes: Select one NWDAF from multiple NWDAF IDs according to a pre-configured local policy; Send an ML Model retrieval request to the selected NWDAF.

5. The method according to claim 2, characterized in that, The step of performing data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtaining the data analysis results corresponding to the expected analysis output information of the AF, specifically includes: Data corresponding to the data type is collected from the data source corresponding to the data source type, and the collected data is analyzed using the target MLModel to obtain the data analysis results corresponding to the analysis output information expected by the AF.

6. A data analysis method based on Network Data Analysis Function (NWDAF), applied to Application Function (AF), characterized in that, The method includes: Send a customized data analysis request to the target NWDAF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured by the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value. The system receives a customized data analysis response message returned by the target NWDAF, which carries the data analysis results. The customized data analysis response message is sent by the target NWDAF after obtaining the corresponding target ML Model based on the AF ID and ML Model ID, performing data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtaining the data analysis results corresponding to the analysis output information expected by the AF. Before sending the customized data analysis request to the target NWDAF, the method further includes: A customized ML Model configuration request is sent to the target NWDAF through the Network Open Function (NEF). The customized ML Model configuration request carries the AF ID, the target ML Model that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target ML Model, ML Model filtering information, and ML Model reporting target. The ML Model filtering information is used to indicate the scope of application of the target ML Model, and the ML Model reporting target is used to indicate the object of data analysis. Receive the customized ML Model configuration response message returned by the target NWDAF through the NEF, the customized ML Model configuration response message carrying the NWDAF ID of the target NWDAF.

7. The method according to claim 6, characterized in that, The data input information required for customized data analysis includes the data types to be collected and the corresponding data source types.

8. The method according to claim 6, characterized in that, The step of sending a customized ML Model configuration request to the target NWDAF via the Network Open Function (NEF) specifically includes: A customized ML Model configuration request is sent to the NEF so that the NEF requests the Network Repository Function (NRF) to discover an NWDAF with Model Training Logic Function (MTLF), and after receiving the target NWDAF returned by the NRF that matches the ML Model filtering information, the NEF sends the customized ML Model configuration request to the target NWDAF.

9. A data analysis device based on Network Data Analysis Function (NWDAF), installed at the target NWDAF, characterized in that, The device includes: The configuration request receiving module is used to receive a customized machine learning model (ML Model) configuration request sent by the application function (AF) through the network open function (NEF). The customized ML Model configuration request carries the AF ID, the target ML Model that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target ML Model, ML Model filtering information, and ML Model reporting target. The ML Model filtering information is used to indicate the scope of application of the target ML Model, and the ML Model reporting target is used to indicate the object of data analysis. The storage module is used to store the target ML Model, the AF ID, ML Model ID, ML Model filtering information, and ML Model reporting target corresponding to the target ML Model, according to the customized ML Model configuration request. A configuration response sending module is configured to return a customized ML Model configuration response message to the AF through the NEF, wherein the customized ML Model configuration response message carries the NWDAF ID of the target NWDAF; The analysis request receiving module is used to receive customized data analysis requests sent by the AF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured in the ML Model required for customized data analysis, the ML Model ID of the ML Model, and the data input information required for customized data analysis and the analysis output information expected by the AF. The Analytics ID is a custom value. The target model acquisition module, connected to the analysis request receiving module, is used to acquire the corresponding target ML Model based on the AF ID and ML ModelID. The data analysis module, connected to the target model acquisition module, is used to perform data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtain the data analysis results corresponding to the analysis output information expected by the AF. The analysis result sending module, connected to the data analysis module, is used to return a customized data analysis response message carrying the data analysis results to the AF.

10. A data analysis device based on Network Data Analysis Function (NWDAF), configured in Application Function (AF), characterized in that, The device includes: The analysis request sending module is used to send a customized data analysis request to the target NWDAF. The customized data analysis request carries the analysis identifier Analytics ID, the AF ID corresponding to the AF, the NWDAF ID of the NWDAF configured by the machine learning model ML Model required for customized data analysis, the ML Model ID of the ML Model, the data input information required for customized data analysis, and the analysis output information expected by the AF. The Analytics ID is a custom value. The analysis result receiving module, connected to the analysis request sending module, is used to receive a customized data analysis response message returned by the target NWDAF, which carries the data analysis results. The customized data analysis response message is sent by the target NWDAF after obtaining the corresponding target ML Model based on the AF ID and ML Model ID, performing data analysis using the target ML Model based on the data input information required for the customized data analysis, and obtaining the data analysis results corresponding to the analysis output information expected by the AF. The device further includes: A configuration request sending module is used to send a customized ML Model configuration request to the target NWDAF through the Network Open Function (NEF). The customized ML Model configuration request carries the AF ID, the target ML Model that the AF wants to configure in the target NWDAF, the ML Model ID corresponding to the target ML Model, ML Model filtering information, and ML Model reporting target. The ML Model filtering information is used to indicate the scope of application of the target ML Model, and the ML Model reporting target is used to indicate the object of data analysis. A configuration response receiving module is configured to receive a customized ML Model configuration response message returned by the target NWDAF through the NEF, wherein the customized ML Model configuration response message carries the NWDAF ID of the target NWDAF.

11. A data analysis device based on Network Data Analysis Function (NWDAF), characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement the NWDAF-based data analysis method as described in any one of claims 1-5, or to implement the NWDAF-based data analysis method as described in any one of claims 6-8.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the data analysis method based on Network Data Analysis Function (NWDAF) as described in any one of claims 1-5, or implements the data analysis method based on NWDAF as described in any one of claims 6-8.

Images