Communication method and apparatus

Abstract

A communication method and apparatus, which relate to the technical field of communications, and are used for improving the acquisition security of model-related data. In the method, an analysis function network element can request a model from a model training function network element by means of a subscription message, and indicate to a data storage network element, by means of an authorization message, that the model training function network element is allowed to obtain model-related data, so that the data storage network element sends the model-related data to the model training function network element, which provides the model, and thus an unauthorized model training function network element or other network elements can be prevented from obtaining the model-related data, thereby improving the acquisition security of the model-related data.

Description

Communication method and device

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to the Chinese patent application filed on August 6, 2023, with application number 202310984335.0 and application name “A communication method and device”, all contents of which are incorporated by reference into this application, and claims priority to the Chinese patent application filed on August 28, 2023, with application number 202311095226.X and application name “A communication method and device”, except for all contents in the application document with the above application number 202310984335.0, which are incorporated by reference into this application. Technical Field

[0003] The present application relates to the field of communication technology, and in particular to a communication method and device. Background Art

[0004] The network data analytics function (NWDAF) element provides data collection, model training, data analysis, and model inference capabilities. It can be used to collect relevant data from network elements, third-party service servers, terminal devices, or network management systems. In scenarios where analysis is performed based on artificial intelligence (AI) or machine learning (ML) models, the NWDAF element can also obtain model-related data and retrain or adjust the original model based on this model-related data.

[0005] Currently, the process of acquiring model-related data lacks a security authentication process, posing security risks.

[0006] Summary of the Invention

[0007] The present application provides a communication method and apparatus for improving the security of obtaining model-related data.

[0008] In a first aspect, a communication method is provided. The method can be implemented by an analysis function network element or a component in the analysis function network element, and the analysis function network element can also be called a communication device or a network device, etc. The analysis function network element is, for example, AnLF. Among them, the components in the present application may include, for example, at least one of a chip, a chip system, a processor, a transceiver, a processing unit, or a transceiver unit. Taking the execution subject as an example, which is a data producer network element, the method can be implemented by the following steps: the analysis function network element sends a subscription message to the model training function network element, and the subscription message is used to request a model; the analysis function network element sends an authorization message to the data storage network element, and the authorization message is used to indicate that the model training function network element is allowed to obtain the data related to the model, and the authorization message includes the identifier of the model training function network element and the index of the data.

[0009] Based on this method, the analysis function network element can request a model from the model training function network element through a subscription message, and indicate to the data storage network element through an authorization message that the model training function network element is allowed to obtain data related to the model, so that the data storage network element sends the model-related data to the model training function network element that provides the model, which can prevent unauthorized model training function network elements or other network elements from obtaining the model-related data, thereby improving the security of obtaining model-related data. The analysis function network element requests a model from the model training function network element, that is, the model training function network element is the network element that provides the model to the analysis function, and the model training function network element needs to train the model based on the model-related data (for example, determine whether to update the model based on the data, or use the data as input to train the model, etc.), and the data is provided by the analysis function network element. The analysis function network element authorizes the model training function network element to obtain data related to the model from the data storage network element. On the one hand, the model training function network element needs to be authorized to obtain data related to the model, which can improve security and prevent unauthorized model training function network elements or other network elements from obtaining data related to the model and further obtaining a model based on the data; on the other hand, the model training function network element has the authority to obtain the data from the data storage network element to train the model, thereby providing the model to the analysis function network element.

[0010] In one possible implementation, the analysis function network element may send the data to the data storage network element and receive the model from the model training function network element, where the model is determined based on the data. For example, before a subscription message is sent, an authorization message is used to indicate that the model training function network element is allowed to obtain the data, so that the model training function network element can obtain the data from the data storage network element and obtain the model based on the data. Thereafter, the analysis function network element may request the model from the model training function network element via a subscription message. Therefore, the data obtained by the analysis function network element authorizing the model training function network element to be used for training the model can also be requested by the analysis function network element from the model training function network element.

[0011] In one possible implementation, the analysis function network element may receive the model from the model training function network element, determine the data based on the model, and send the data to the data storage network element. Therefore, the data authorized by the analysis function network element to the model training function network element may be obtained based on the model from the model training function network element. The analysis function network element may also store the data in the data storage network element, allowing the data storage network element to provide the data to the model training function network element. Optionally, in this implementation, the data may be used by the model training function network element to determine whether to update the model.

[0012] In one possible implementation, the analysis function network element may also send the analysis identifier to the network storage function network element, and the analysis function network element receives the identifier of the model training function network element from the network storage function network element. The identifier of the model training function network element is determined based on the analysis identifier. In various implementations of the present application, the analysis identifier can be used to determine the model training function network element together with the manufacturer identifier and / or interoperability identifier, thereby achieving accurate indication of the model training function network element. In addition, this implementation can support the analysis function network element to efficiently obtain the manufacturer identifier and / or interoperability identifier of the model training function network element through the analysis identifier.

[0013] In a second aspect, a communication method is provided. The method can be implemented by a data storage network element or a component in the data storage network element, and the data storage network element can also be called a communication device or a network device, etc. The data storage network element is, for example, an ADRF. Taking the execution subject as an example, the method can be implemented by the following steps: the data storage network element receives an authorization message, the authorization message is used to indicate that the model training function network element is allowed to obtain model-related data, and the authorization message includes the identifier of the model training function network element and the index of the data; the data storage network element receives a data request from the model training function network element, and the data request is used to request the data; when the model training function network element is allowed to obtain the model-related data, the data storage network element sends the data to the model training function network element.

[0014] Optionally, a model training function network element may be used to train the model.

[0015] In one possible implementation, the list of permitted network function consumers for the data in the authorization message includes an identifier of the model training function network element; if the list of permitted network function consumers for the data includes the identifier of the model training function network element, the data storage network element sends the data to the model training function network element. This list may also be referred to as a list of permitted consumers for the data or a list of permitted data consumers for the data.

[0016] In a possible implementation, the data storage network element may store a list of network function consumers allowed for the data.

[0017] In a possible implementation, the authorization message comes from an analysis function network element, and the method further includes: the data storage network element receiving the data from the analysis function network element and storing the data.

[0018] In a possible implementation of the first aspect or the second aspect, the identifier of the model training function network element includes one or more of the device identifier of the model training function network element, the manufacturer identifier of the model training function network element, and the interoperability identifier of the model training function network element.

[0019] Based on this implementation, the interoperability identifier corresponds to the manufacturer identifier, so the model training functional network elements that are allowed to obtain data can be indicated at device granularity or manufacturer granularity.

[0020] In a possible implementation manner of the first aspect or the second aspect, the list of allowed network function consumers for the data in the authorization message includes an identifier of the model training function network element.

[0021] In a possible implementation of the first aspect or the second aspect, the identifier of the model training function network element includes a vendor identifier or an interoperability identifier of the model training function network element, the subscription message includes the analysis identifier, the authorization message also includes the analysis identifier, and the data request includes the analysis identifier. The analysis identifier can be used to identify the model training function network element together with the vendor identifier and / or interoperability identifier, thereby enabling accurate indication of the model training function network element.

[0022] In one possible implementation of the first or second aspect, the data request includes an identifier of the analysis function network element and / or an index of the data. In other words, the data request includes an index of the data, and the index of the data may include an identifier of the analysis function network element. Therefore, the data may be identified based on the identifier of the analysis function network element and the index of the data to avoid index duplication.

[0023] In a possible implementation of the first aspect or the second aspect, the subscription message includes an index of the data.

[0024] In a possible implementation of the first aspect or the second aspect, the authorization message also includes at least one of an identifier of the analysis function network element and an authentication indication, and the authentication indication is used to indicate that the model training function network element that obtains the data needs authorization, or in other words, the authentication indication is used to indicate that authorization is required to obtain the data. Based on this implementation, the analysis function network element can accurately indicate the data that requires verification of the authority of the model training function network element through the authorization message, for example, through the authentication indication, indicate to the data storage network element that the data requires authorization to be read.

[0025] In a possible implementation of the first aspect or the second aspect, the index of the data may be used to obtain data related to the model.

[0026] The beneficial effects of each possible implementation method in the above second aspect can be found in the description of the beneficial effects of the first aspect and its various aspects.

[0027] In a third aspect, a communication method is provided. The method can be implemented by an analysis function network element or a component in the analysis function network element, and the analysis function network element can also be referred to as a communication device or a network device, etc. The analysis function network element is, for example, AnLF. Among them, the components in the present application may include, for example, at least one of a chip, a chip system, a processor, a transceiver, a processing unit, or a transceiver unit. Taking the execution subject as an analysis function network element as an example, the method can be implemented by the following steps: the analysis function network element obtains the identifier of the model training function network element; the analysis function network element sends a subscription message to the model training function network element, the subscription message is used to request a model, the subscription message includes a first token, the first token is used to provide authorization for the model training function network element to obtain data related to the model, and the first token includes the identifier of the analysis function network element, the identifier of the model training function network element, and the index of the data.

[0028] Based on this method, the analysis function network element can send a first token to the model training function network element and request a model from the model training function network element. The first token can be used to provide authorization for obtaining data related to the model. Therefore, the model training function network element can request model-related data from the data storage network element based on the first token, so that the data storage network element sends the model-related data to the model training function network element that provides the model, which can prevent unauthorized model training function network elements or other network elements from obtaining model-related data, thereby improving the security of obtaining model-related data.

[0029] In one possible implementation, the analysis function network element may send the data to the data storage network element and receive the model from the model training function network element, where the model is determined based on the data. For example, before a subscription message is sent, an authorization message is used to indicate that the model training function network element is allowed to obtain the data, so that the model training function network element can obtain the data from the data storage network element and obtain the model based on the data. Thereafter, the analysis function network element may request the model from the model training function network element via a subscription message. Therefore, the data obtained by the analysis function network element authorizing the model training function network element to be used for training the model can also be requested by the analysis function network element from the model training function network element.

[0030] In one possible implementation, the analysis function network element may receive the model from the model training function network element, determine the data based on the model, and send the data to the data storage network element. Therefore, the data authorized by the analysis function network element to the model training function network element may be obtained based on the model from the model training function network element. The analysis function network element may also store the data in the data storage network element, allowing the data storage network element to provide the data to the model training function network element. Optionally, in this implementation, the data may be used by the model training function network element to determine whether to update the model.

[0031] In one possible implementation, the analysis function network element may further send an authorization message to the data storage network element, where the authorization message includes at least one of an identifier of the analysis function network element, an index of the data, and an authentication indication, where the authentication indication is used to indicate that authorization is required to obtain the data. The authentication indication is used to indicate that authorization is required to obtain the data.

[0032] In a possible implementation, the analysis function network element may also receive a token update request from the model training function network element, where the token update request is used to update the first token.

[0033] In a fourth aspect, a communication method is provided. The method can be implemented by a data storage network element or a component in the data storage network element, and the data storage network element can also be referred to as a communication device or network device, etc. The data storage network element is, for example, an ADRF. Taking the execution subject as an example where the data storage network element is an example, the method can be implemented by the following steps: the data storage network element receives a data request from the model training function network element, the data request is used to request the data, and the data request includes the first token; the data storage network element sends the data to the model training function network element according to the first token. For various technical terms, please refer to the description in the fourth aspect.

[0034] In one possible implementation, the data request includes an index of the data. Upon verifying that the index of the data in the data request matches the index of the data in the first token, the data storage network element sends the data to the model training function network element. Therefore, if the index of the data in the first token is consistent with or matches the index of the data requested by the model analysis function network element, the data can be sent to the model analysis function network element, thereby improving the security of the data acquisition process.

[0035] In one possible implementation, the data request also includes an identifier of the analysis function network element, and when the identifier of the analysis function network element in the data request is consistent with the identifier of the analysis function network element in the first token, the data storage network element sends the data to the model training function network element. The identifier of the analysis function network element carried in the request message of the model analysis function network element can be used to indicate the producer of the data requested by the model analysis function network element. Therefore, when the identifier of the analysis function network element in the first token is consistent with or matches the identifier of the producer of the data requested by the model analysis function network element, data can be sent to the model analysis function network element to improve the security of the data acquisition process.

[0036] In one possible implementation, the data storage network element may also receive an authorization message, wherein the authorization message includes an identifier of the analysis function network element, an index of the data, and at least one of an authentication indication, wherein the authentication indication is used to indicate that authorization is required to obtain the data.

[0037] In a possible implementation, the data storage network element may specifically send the data to the model training function network element when the first token includes the identifier of the analysis function network element in the authorization message and the index of the data.

[0038] In a possible implementation, the authorization message comes from an analysis function network element, and the method further includes: the data storage network element receiving the data from the analysis function network element and storing the data.

[0039] In a fifth aspect, a communication method is provided. The method can be implemented by a model training function network element or a component in the model training function network element. The model training function network element can also be referred to as a communication device or network equipment, etc. The model training function network element is, for example, an MTLF. Taking the execution subject as an example, the method can be implemented by the following steps: the model training function network element receives a subscription message from the analysis function network element, the subscription message is used to request a model, the subscription message includes a first token, the first token is used to authorize the model training function network element to obtain data related to the model, the first token includes the identifier of the analysis function network element, the identifier of the model training function network element, and the index of the data; the model training function network element sends a data request to the data storage network element, the data request is used to request the data, and the data request includes the first token; the model training function network element receives the data from the data storage network element.

[0040] In one possible implementation, the model training function network element may also determine the model based on the data and send the model to the analysis function network element. For example, before sending a subscription message, the analysis function network element may indicate through an authorization message that it is allowed to obtain data from the model training function network element, so that the model training function network element can obtain the data from the data storage network element and obtain the model based on the data. Thereafter, the analysis function network element may request the model from the model training function network element through a subscription message. Therefore, the analysis function network element authorizes the model training function network element to obtain data that can be used to train the model, and the analysis function network element may also request the model from the model training function network element.

[0041] In one possible implementation, the model training functional network element sends the model to the analysis functional network element, and the model is used to determine the data. Therefore, the data authorized by the analysis functional network element to the model training functional network element can be obtained based on the model from the model training functional network element. The analysis functional network element can also store the data in the data storage network element, so that the data storage network element can provide the data to the model training functional network element. Optionally, in this implementation, the model training functional network element determines to update the model based on the data and sends the updated model to the analysis functional network element.

[0042] In a possible implementation, the model training function network element may also send a token update request to the analysis function network element, where the token update request is used to update the first token.

[0043] In a possible implementation of the third aspect, the fourth aspect, or the fifth aspect, the first token further includes validity period information of the first token.

[0044] In a possible implementation of the third aspect, the fourth aspect, or the fifth aspect, the subscription message may further include the analysis identifier.

[0045] In a possible implementation of the fourth aspect or the fifth aspect, the data request may also include at least one of an analysis identifier, an identifier of the analysis function network element, an index of the data, and an authentication indication.

[0046] In a possible implementation of the third aspect, the fourth aspect, or the fifth aspect, the index of the data may be used to obtain data related to the model.

[0047] The beneficial effects of each possible implementation method in the fourth and fifth aspects above can be found in the description of the beneficial effects of the third aspect and its various aspects.

[0048] In a sixth aspect, a communication method is provided. The method can be implemented by a network storage function network element or a component in the network storage function network element. The network storage function network element can also be referred to as a communication device or network equipment, etc. The network storage function network element is, for example, an NRF. Among them, the components in the present application may include, for example, at least one of a chip, a chip system, a processor, a transceiver, a processing unit, or a transceiver unit. Taking the execution subject as an example, the method can be implemented by the following steps: the network storage function network element receives a token request from the model training function network element, and the token request includes an identifier of the analysis function network element; when the model training function network element is allowed to obtain the data provided by the analysis function network element, the network storage function network element sends a second token to the model training function network element, and the second token is used to provide authorization for the model training function network element to obtain the data, and the second token includes the identifier of the analysis function network element and the identifier of the model training function network element. Optionally, the model training function network element can be used to train the model, and the analysis function network element can be used to provide model-related data.

[0049] Based on this method, the network storage function network element can receive a token request from the model training function network element, and provide a second token to the model training function network element while allowing the model training function network element to obtain model-related data. The second token can be used to provide authorization for the model training function network element to obtain the model-related data. Therefore, the model training function network element can request model-related data from the data storage network element based on the second token, so that the data storage network element sends the model-related data to the model training function network element that provides the model, which can prevent unauthorized model training function network elements or other network elements from obtaining the model-related data, thereby improving the security of obtaining model-related data.

[0050] In one possible implementation, the network storage function network element may receive a list of allowed network function consumers for the data from the analysis function network element, the list including the identifier of the model training function network element, wherein the model training function network element is allowed to obtain the data provided by the analysis function network element if the list includes the identifier of the model training function network element.

[0051] In one possible implementation, the network storage function network element may receive an analysis identifier from the analysis function network element, and the model is used to execute the analysis service corresponding to the analysis identifier; determine the model training function network element that supports the analysis service corresponding to the analysis identifier based on the analysis identifier, and send the identifier of the model training function network element to the analysis function network element.

[0052] In the seventh aspect, a communication method is provided. The method can be implemented by an analysis function network element or a component in the analysis function network element, and the analysis function network element can also be called a communication device or a network device, etc. The analysis function network element is, for example, AnLF. Among them, the components in this application may include, for example, at least one of a chip, a chip system, a processor, a transceiver, a processing unit, or a transceiver unit. Taking the execution subject as an analysis function network element as an example, the method can be implemented by the following steps: the analysis function network element can send a list of allowed network function consumers for the data to the network storage function network element, and the list includes an identifier of the model training function network element. Optionally, the model training function network element can be used to train the model, and the analysis function network element can be used to provide model-related data.

[0053] In one possible implementation, after sending the list, the analysis function network element may receive the model from the model training function network element, where the model is determined based on the data. Specifically, the analysis function network element first authorizes the model training function network element to obtain the data, and then the model training function network element obtains a model based on the data and provides the model to the analysis function network element.

[0054] In one possible implementation, the analysis function network element may send the data to the data storage network element, receive the model from the model training function network element, determine the data based on the model, and send the data to the data storage network element. Specifically, the model training function network element first obtains a model based on the data and provides the model to the analysis function network element, which then obtains model-related data based on the model. Optionally, in this implementation, the data may be used by the model training function network element to determine whether to update the model.

[0055] In one possible implementation, the analysis function network element may send an analysis identifier to the network storage function network element, and the model is used to execute the analysis service corresponding to the analysis identifier; the analysis function network element receives the identifier of the model training function network element from the network storage function network element, and the identifier of the model training function network element is determined based on the analysis identifier.

[0056] In an eighth aspect, a communication method is provided. The method can be implemented by a model training function network element or a component in the model training function network element, and the model training function network element can also be referred to as a communication device or a network device, etc. The model training function network element is, for example, an MTLF. Taking the execution subject as an example, the method can be implemented by the following steps: the model training function network element sends a token request to the network storage function network element, and the token request includes the identifier of the analysis function network element; the model training function network element receives a second token from the network storage function network element, and the second token includes the identifier of the analysis function network element and the identifier of the model training function network element; the model training function network element sends a data request to the data storage network element, and the data request is used to request the model-related data provided by the analysis function network element, and the data request includes the second token; the model training function network element receives the data from the data storage network element.

[0057] In a possible implementation, the model training function network element may also determine the model based on the data and send the model to the analysis function network element.

[0058] In one possible implementation, the model training function network element sends the model to the analysis function network element, and the model is used to determine the data. Optionally, in this implementation, the model training function network element determines to update the model based on the data and sends the updated model to the analysis function network element.

[0059] In a possible implementation of the sixth, seventh, or eighth aspects, the second token further includes the analysis identifier and / or the data storage device identifier. The analysis identifier can be used together with the vendor identifier and / or the interoperability identifier to identify the model training function network element. The data storage device identifier can be used to ensure that the model training function network element can only obtain data from the data storage network element specified by the analysis function network element.

[0060] In a possible implementation manner of the sixth, seventh, or eighth aspect, the token request further includes the analysis identifier and / or an identifier of the data storage device.

[0061] In a possible implementation of the sixth, seventh or eighth aspect, the identifier of the model training function network element includes one or more of the following: the device identifier of the model training function network element, the manufacturer identifier of the model training function network element, and the interoperability identifier of the model training function network element.

[0062] In a ninth aspect, a communication device is provided. The device can implement the method described in any one of the first to eighth aspects above and any possible designs thereof. The device has the functions of the above-mentioned analysis function network element, data storage network element, model training function network element, or network storage function network element. The device is, for example, an analysis function network element, a data storage network element, a model training function network element, or a network storage function network element, or is a functional module in the analysis function network element, the data storage network element, the model training function network element, or the network storage function network element.

[0063] In an optional implementation, the device may include a module corresponding to the method / operation / step / action described in any one of the first to eighth aspects and any possible designs thereof. The module may be a hardware circuit, software, or a combination of a hardware circuit and software. In an optional implementation, the device includes a processing unit (sometimes also referred to as a processing module) and a communication unit (sometimes also referred to as a transceiver module, a communication module, etc.). The transceiver unit is capable of implementing a sending function and a receiving function. When the transceiver unit implements the sending function, it may be referred to as a sending unit (sometimes also referred to as a sending module). When the transceiver unit implements the receiving function, it may be referred to as a receiving unit (sometimes also referred to as a receiving module). The sending unit and the receiving unit may be the same functional module, which is called a transceiver unit, and the functional module is capable of implementing a sending function and a receiving function; or, the sending unit and the receiving unit may be different functional modules, and the transceiver unit is a general term for these functional modules.

[0064] Exemplarily, when the apparatus is used to execute the method described in any one of the first aspect to the eighth aspect and any possible design thereof, the apparatus may include a communication unit and a processing unit.

[0065] In the tenth aspect, an embodiment of the present application also provides a communication device, including a processor for executing a computer program (or computer executable instructions) stored in a memory, so that when the computer program (or computer executable instructions) is executed, the device executes the method described in any one of the first to eighth aspects and any possible designs thereof.

[0066] In one possible implementation, the processor and memory are integrated;

[0067] In another possible implementation, the memory is located outside the communication device.

[0068] The communication device also includes a communication interface, which is used for the communication device to communicate with other devices, such as sending or receiving data and / or signals. Exemplarily, the communication interface can be a transceiver, circuit, bus, module or other type of communication interface.

[0069] In the eleventh aspect, a computer-readable storage medium is provided, which is used to store computer programs or instructions, which, when executed, enables the method described in any one of the first to eighth aspects and any possible designs thereof to be implemented.

[0070] In a twelfth aspect, a computer program product comprising instructions is provided, which, when executed on a computer, enables the method described in any one of the first to eighth aspects and any possible designs thereof to be implemented.

[0071] In the thirteenth aspect, an embodiment of the present application also provides a communication device for executing the method described in any one of the above-mentioned aspects from the first to the eighth aspect and any possible design thereof.

[0072] In the fourteenth aspect, a chip system is provided, which includes a logic circuit (or it can be understood that the chip system includes a processor, and the processor may include a logic circuit, etc.), and may also include an input and output interface. The input and output interface can be used to input messages and also to output messages. The input and output interfaces can be the same interface, that is, the same interface can implement both the sending function and the receiving function; or, the input and output interfaces include an input interface and an output interface, the input interface is used to implement the receiving function, that is, for receiving messages; the output interface is used to implement the sending function, that is, for sending messages. The logic circuit can be used to perform the operations other than the sending and receiving functions in the method described in any of the first to eighth aspects above and any possible designs thereof; the logic circuit can also be used to transmit messages to the input and output interface, or receive messages from other communication devices from the input and output interface. The chip system can be used to implement the method described in any of the first to eighth aspects above and any possible designs thereof. The chip system can be composed of a chip, or it can include a chip and other discrete devices.

[0073] Optionally, the chip system may further include a memory, which may be used to store instructions, and the logic circuit may call the instructions stored in the memory to implement corresponding functions.

[0074] In a fifteenth aspect, a communication system is provided, which may include at least one communication device for implementing the method shown in any of the above aspects.

[0075] For example, the communication system may include an analysis function network element for implementing the method shown in the first aspect and any possible implementation thereof, and a data storage network element for implementing the method shown in the second aspect and any possible implementation thereof.

[0076] For example, the communication system may include an analysis function network element for implementing the method shown in the third aspect and any possible implementation thereof, a data storage network element for implementing the method shown in the fourth aspect and any possible implementation thereof, and a model training function network element for implementing the method shown in the fifth aspect and any possible implementation thereof.

[0077] For example, the communication system may include a network storage function network element for implementing the method shown in the sixth aspect and any possible implementation thereof, an analysis function network element for implementing the method shown in the seventh aspect and any possible implementation thereof, and a model training function network element for implementing the method shown in the eighth aspect and any possible implementation thereof.

[0078] The technical effects brought about by the above second to ninth aspects can be found in the description of the above first aspect and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] FIG1 is a schematic diagram of the architecture of a wireless communication system provided in an embodiment of the present application;

[0080] FIG2 is a schematic diagram of the architecture of another wireless communication system provided in an embodiment of the present application;

[0081] FIG3 is a flow chart of a communication method provided in an embodiment of the present application;

[0082] FIG4 is a flow chart of another communication method provided in an embodiment of the present application;

[0083] FIG5 is a flow chart of another communication method provided in an embodiment of the present application;

[0084] FIG6 is a flow chart of another communication method provided in an embodiment of the present application;

[0085] FIG7 is a flow chart of another communication method provided in an embodiment of the present application;

[0086] FIG8 is a flow chart of another communication method provided in an embodiment of the present application;

[0087] FIG9 is a schematic structural diagram of a communication device provided in an embodiment of the present application;

[0088] FIG10 is a schematic diagram of the structure of another communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0089] Figure 1 is a schematic diagram of a 5G network architecture based on a service-oriented architecture. The 5G network architecture shown in Figure 1 includes terminal devices, access network devices, and core network devices. Terminal devices access the data network (DN) through the access network and core network devices. Among them, the core network equipment includes a variety of network functions (NF) or network elements, for example, including some or all of the following network elements: unified data management (UDM) network element, unified data repository (UDR) network element, network exposure function (NEF) network element (not shown in the figure), application function (AF) network element, policy control function (PCF) network element, access and mobility management function (AMF) network element, session management function (SMF) network element, user plane function (UPF) network element, network data analytics function (NWDAF) network element, network repository function (NRF) network element (not shown in the figure), and location management function (LMF) network element (not shown in the figure).

[0090] The access network device may be a radio access network (RAN) device. For example: a base station, an evolved NodeB (eNodeB), a transmission reception point (TRP), a next generation NodeB (gNB) in a 5G mobile communication system, a next generation base station in a sixth generation (6G) mobile communication system, a base station in a future mobile communication system, or an access node in a wireless fidelity (WiFi) system, etc.; it may also be a module or unit that performs part of the functions of a base station, for example, a centralized unit (CU) or a distributed unit (DU). The radio access network device may be a macro base station, a micro base station or an indoor station, a relay node or a donor node, etc. The embodiments of the present application do not limit the specific technology and specific device form adopted by the radio access network device.

[0091] Terminal devices can be user equipment (UE), mobile stations, mobile terminals, etc. Terminal devices can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), the Internet of Things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, and smart cities. Terminal devices can be mobile phones, tablets, computers with wireless transceiver capabilities, wearable devices, vehicles, urban air vehicles (such as drones and helicopters), ships, robots, robotic arms, smart home devices, etc.

[0092] Access network equipment and terminal devices can be fixed or mobile. They can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; on water; or in the air on aircraft, balloons, and satellites. The embodiments of this application do not limit the application scenarios of access network equipment and terminal devices.

[0093] The following describes the core network equipment involved in this application:

[0094] The AMF network element includes functions such as mobility management and access authentication / authorization. In addition, it is also responsible for transmitting user policies between terminal devices and PCF. The SMF network element includes functions such as session management, execution of control policies issued by PCF, selection of UPF, and allocation of Internet Protocol (IP) addresses for terminal devices. The UPF network element, as an interface with the data network, includes functions such as user plane data forwarding, session / flow-level billing statistics, and bandwidth limitation. The UDM network element includes functions such as management of subscription data and user access authorization. The UDR network element includes functions for accessing subscription data, policy data, application data, and other types of data. The NEF network element is used to support the opening of capabilities and events. The AF network element transmits the requirements of the application side to the network side, such as QoS requirements or user status event subscriptions. The AF can be a third-party functional entity or an application server deployed by the operator. The PCF network element includes policy control functions such as billing at the session and service flow level, QoS bandwidth guarantee and mobility management, and terminal device policy decision-making. NRF network elements can be used to provide network element discovery functions, providing network element information corresponding to the network element type based on requests from other network elements. NRF network elements also provide network element management services, such as network element registration, update, deregistration, and network element status subscription and push.

[0095] The NWDAF network element is primarily used to collect data (including one or more of terminal device data, access network device data, core network element data, and third-party application device data). This data can be the terminal device, access network device, core network element, or third-party application device's own data, or the terminal device's data on the access network device, core network element, or third-party application device. The collected data is then analyzed and the analysis results are output for use by the network, network management equipment, and application execution policy decisions. The NWDAF can use machine learning models for data analysis. Release 17 of the 3rd Generation Partnership Project (3GPP) separates the training and inference functions of the NWDAF. An NWDAF can support only model training, only data inference, or both model training and data inference. An NWDAF that supports model training can also be referred to as a training NWDAF, or an NWDAF that supports the model training logical function (MTLF) (abbreviated as MTLF). The training NWDAF can perform model training based on the acquired data to obtain a trained model. An NWDAF that supports data reasoning function can also be called a reasoning NWDAF, or a NWDAF that supports analytics logical function (AnLF) (abbreviated as AnLF). The reasoning NWDAF can input input data into the trained model to obtain analysis results or reasoning data. In the embodiment of the present application, the training NWDAF refers to an NWDAF that at least supports the model training function. As a possible implementation method, the training NWDAF can also support the data reasoning function. The reasoning NWDAF refers to an NWDAF that at least supports the data reasoning function. As a possible implementation method, the reasoning NWDAF can also support the model training function. If an NWDAF supports both model training function and data reasoning function, the NWDAF can be called a training NWDAF, a reasoning NWDAF, a training reasoning NWDAF, or an NWDAF. In the embodiment of the present application, an NWDAF can be a separate network element, or it can be set up together with other network elements, for example, the NWDAF is set up in a PCF network element or an AMF network element.

[0096] The analytics data repository functional (ADRF) network element may be referred to as ADRF. The ADRF may be used to store model-related data. The data may be generated by the AnLF. The ADRF may provide model-related data to the MTLF upon request. The ADRF is not shown in Figures 1 or 2.

[0097] The LMF network element is used to manage the location information of terminal devices. It can calculate or verify the location of terminal devices and / or estimate the speed of terminal devices, and provide the estimated accuracy. The LMF network element can receive the location request of the terminal device from the AMF network element through the Nlmf interface. The granularity of the terminal location calculated by the LMF network element can be one or more of longitude, latitude, altitude, cell, and global positioning system (GPS).

[0098] A DN is a network located outside of a carrier network. A carrier network can connect to multiple DNs, and a variety of services can be deployed on the DN, providing data and / or voice services to terminal devices. For example, a DN is the private network of a smart factory. Sensors installed in the workshop can be terminal devices. The DN houses a sensor control server, which provides services to the sensors. Sensors can communicate with the control server, receive instructions from the control server, and transmit collected sensor data to the control server based on the instructions. Another example is a DN that is a company's internal office network. An employee's mobile phone or computer can be a terminal device, allowing them to access information and data resources on the company's internal office network.

[0099] It can be understood that the above network elements are examples of one implementation method, and this application does not exclude the existence of network elements or devices with the above network element functions in 6G or newer wireless communication systems that have other names or other forms.

[0100] In Figure 1, Nudr, Npcf, Namf, Nudm, Nsmf, Naf, and Nnwdaf are service-oriented interfaces provided by the UDR, PCF, AMF, UDM, SMF, AF, and NWDAF, respectively, and are used to invoke corresponding service-oriented operations. N1, N2, N3, N4, and N6 are interface serial numbers, and their meanings are as follows:

[0101] 1) N1: The interface between the AMF network element and the terminal device, which can be used to transmit non-access stratum (NAS) signaling (such as QoS rules from the AMF network element) to the terminal device.

[0102] 2) N2: The interface between the AMF network element and the access network equipment, which can be used to transmit radio bearer control information from the core network side to the access network equipment.

[0103] 3) N3: The interface between the access network equipment and the UPF network element, mainly used to transmit uplink and downlink user plane data between the access network equipment and the UPF network element.

[0104] 4) N4: The interface between the SMF network element and the UPF network element, which can be used to transmit information between the control plane and the user plane, including controlling the issuance of forwarding rules, QoS rules, traffic statistics rules, etc. for the user plane and reporting information on the user plane.

[0105] 5) N6: Interface between UPF network element and DN, used to transmit uplink and downlink user data flows between UP network element F and DN.

[0106] In addition, in the architecture shown in Figure 1, Nadrf can serve as a service-oriented interface for ADRF.

[0107] Figure 2 is a schematic diagram of a 5G network architecture based on point-to-point interfaces. The functions of the network elements in Figure 2 can be referred to as those of the corresponding network elements in Figure 1 and will not be repeated here. The main difference between Figure 2 and Figure 1 is that the interfaces between the control plane network elements in Figure 1 are service-oriented interfaces, while the interfaces between the control plane network elements in Figure 2 are point-to-point interfaces.

[0108] In the architecture shown in Figure 2, the interface names and functions between the various network elements of the core network are as follows:

[0109] 1) N5: The interface between the AF network element and the PCF network element, which can be used to issue application service requests and report network events.

[0110] 2) N7: The interface between the PCF network element and the SMF network element, which can be used to issue protocol data unit (PDU) session granularity and service data flow granularity control strategy.

[0111] 3) N8: The interface between the AMF network element and the UDM network element, which can be used by the AMF network element to obtain access and mobility management related contract data and authentication data from the UDM network element, and the AMF network element to register the current mobility management related information of the terminal device to the UDM network element.

[0112] 4) N9: User plane interface between UPF network elements, used to transmit uplink and downlink user data flows between UPF network elements.

[0113] 5) N10: The interface between the SMF network element and the UDM network element, which can be used by the SMF network element to obtain session management related contract data from the UDM network element, and the SMF network element to register the current session related information of the terminal device to the UDM network element.

[0114] 6) N11: The interface between the SMF network element and the AMF network element, which can be used to transmit PDU session tunnel information between the access network device and the UPF network element, transmit control messages sent to the terminal device, transmit wireless resource control information sent to the access network device, etc.

[0115] 7) N15: The interface between the PCF network element and the AMF network element, which can be used to issue terminal device policies and access control related policies.

[0116] 8) N23: The interface between the PCF network element and the NWDAF network element. The NWDAF network element can collect data on the PCF network element through this interface. It should be noted that the NWDAF network element can also have interfaces with other devices (such as AMF network elements, UPF network elements, access network equipment, terminal equipment, etc.), which are not fully shown in the figure.

[0117] 9) N35: The interface between the UDM network element and the UDR network element, which can be used by the UDM network element to obtain user contract data information from the UDR network element.

[0118] 10) N36: Interface between PCF network element and UDR network element, which can be used by PCF network element to obtain policy-related contract data and application data-related information from UDR network element.

[0119] It is understood that the above-mentioned network element or function can be a network element in a hardware device, a software function running on dedicated hardware, or a virtualized function instantiated on a platform (e.g., a cloud platform). As a possible implementation method, the above-mentioned network element or function can be implemented by a single device, or can be implemented by multiple devices together, or can be a functional module within a single device, which is not specifically limited in the embodiments of the present application.

[0120] As an implementation method, the data analysis network element in the embodiment of the present application can be the above-mentioned NWDAF network element, or it can be a network element having the functions of the above-mentioned NWDAF network element in future communications such as 6G networks. For ease of explanation, the following description is based on the example that the data analysis network element is an NWDAF network element, and the NWDAF network element is referred to as NWDAF for short.

[0121] In this application, AnLF can be used to perform inference, derive analysis information, and expose analysis services. Analysis can refer to the export of statistical information and / or predictions based on requests from analysis consumers. MTLF can be used to train machine learning (ML) models and expose new training services, for example, providing trained ML models to AnLF. In this application, for the process of obtaining model-related data, AnLF can serve as a data producer network element, denoted as NFp; MTLF can serve as a data consumer network element, denoted as NFc.

[0122] Currently, AnLF can request a model from MTLF through the model subscription (MLModelProvision_Subscribe) service or message. The model can be obtained by MTLF through training based on the relevant data of the model. In addition, AnLF can use a data set tag to specify a data set, which can contain data related to a model. Therefore, the data set tag can be used to mark a data set, that is, to mark the data related to the model. The data set tag can also be understood as the index of the data set or the index of the data. MTLF can act as a consumer network element of model-related data to obtain data stored in the data store by the data producer network element, where the data producer network element can be AnLF and the data store network element can be a network element for storing data such as the analytics data repository functional (ADRF).

[0123] Specifically, model-related data may include any data from the model input data, model output data (e.g., prediction results or prediction data), model true value data, and deviation data during prediction from ADRF. Among them, the model input data can be used to train and obtain the model. The model output data is, for example, the prediction result obtained based on the input data, or the model prediction result corresponding to the input data. The model true value data can be used to represent the true value of the output data corresponding to the model input data, and can also be called the true result of the input data. The deviation data during prediction can identify the deviation between the prediction data and the true value data. The model output data and the model true value data can be used for monitoring the accuracy of the model and retraining and / or re-providing relevance. That is, the model can be obtained based on the model training data, wherein, when the model is obtained by the first training, the model can be obtained based on the model input data and the model true value data. In addition, the model can also be obtained by retraining on the basis of the historical model based on the data related to the model, and the historical model is an existing model. Among them, the prediction data can be the output result of the model.

[0124] Currently, the process for acquiring model-related data is the same as that for acquiring non-model-related data. Therefore, model-related data is not protected. As long as the MTLF provides dataset labels, model-related data can be obtained from the ADRF. Therefore, if this data is maliciously traversed and obtained by an attacker, they may use this data to retrain a model with similar functionality, thereby completing a model theft attack, thus posing a security risk.

[0125] The present application provides a communication method, in which the model training function network element that is authorized by the analysis function network element to provide the model obtains model-related data. Therefore, the data storage network element can verify whether the model training function network element has the authority to obtain the model-related data before providing the model-related data, so as to avoid the model training function network element that does not obtain the authority to obtain the model-related data, resulting in model leakage, thereby improving the security of the model-related data. Among them, the data storage network element may include ADRF or other network elements or devices for storing model-related data. That is, the model training function network element can be used to provide the model to the analysis function network element, and obtain model-related data based on the authorization of the analysis function network element. The model training function network element, for example, includes MTLF or other NWDAF. The analysis function network element can obtain the model from the model training function network element and authorize the model training function network element to obtain model-related data. The analysis function network element, for example, includes AnLF or other NWDAF.

[0126] In the following, an example is given in which the data storage network element is ADRF, the model training function network element is MTLF, and the analysis function network element is AnLF.

[0127] It is understood that in various embodiments of the present application, AnLF can be used to generate model-related data and store the model-related data in ADRF, and MTLF can obtain the model-related data from ADRF. It can be seen that for model-related data, this data is generated by AnLF, so AnLF can serve as a data producer network element. In addition, model-related data can be used by MTLF to train or update the model, so MTLF can serve as a data consumer network element. Among them, updating the model can include retraining the model, and can also include re-obtaining the model based on new input data.

[0128] It is understood that some or all of the model-related data can be generated directly by AnLF, such as the output data of the model, which is generated when AnLF uses the model. Some or all of the model-related data can also be obtained by AnLF from other network elements or devices, such as some or all of the model's input data is collected by AnLF from other network elements, such as AMF or UPF. In various embodiments of the present application, AnLF stores the model-related data in ADRF, so AnLF can be considered as the data producer of the model-related data.

[0129] As an example of an implementation method, the model can be obtained by MTLF based on model-related data training. For example, AnLF generates or collects model-related data from other network elements and sends the data to ADRF. The corresponding ADRF stores the data. AnLF can also authorize MTLF to obtain the data from ADRF. After obtaining the data, MTLF can perform model training to obtain a model and send the model to AnLF. Therefore, at this time, the model-related data may include the input data of the model or the model's true value data. At this time, AnLF can first authorize MTLF to obtain model-related data, and then AnLF can send a subscription message to MTLF to request MTLF to provide AnLF with the model obtained by training based on the data.

[0130] As another example of an implementation, AnLF can first send a subscription message to MTLF, causing MTLF to provide a model to AnLF. AnLF can then obtain model-related data based on the model and store the data in ADRF. AnLF can also authorize MTLF to obtain model-related data, allowing MTLF to obtain model-related data from ADRF based on the authorization and determine whether to update the model based on the model-related data. If the model needs to be updated, the model can be updated based on the model-related data to obtain an updated model. In this case, the model-related data, for example, includes model output data and / or model true value data of the model. The model output data and / or model true value data can be used by MTLF to determine whether to update the model previously provided to AnLF.

[0131] The above two examples can also be implemented in combination. For example, AnLF first sends data to ADRF and authorizes MTLF to obtain the data from ADRF. MTLF can then train a model based on the data and provide the trained model to AnLF. Furthermore, AnLF can generate new data based on the model and send the generated data to ADRF. MTLF can also obtain new data from ADRF based on AnLF's authorization and determine whether to update the model previously provided to AnLF based on the new data. If the model needs to be updated, MTLF can retrain the model based on the new data and send the updated model to AnLF.

[0132] The communication method provided by this application is introduced below with reference to a flowchart.

[0133] As an embodiment, the AnLF may send an authorization message to the ADRF, and the authorization message may be used to indicate that the MTLF is allowed to obtain model-related data. As shown in FIG3 , an implementation of this embodiment may include the following steps:

[0134] S101: AnLF sends a subscription message to MTLF. The subscription message can be used to request a model.

[0135] It should be understood that in this application, the subscription message can also be referred to as a model request message, etc.

[0136] As described in this application, AnLF can request a model from MTLF and authorize the MTLF to obtain data related to the model.

[0137] The AnLF may obtain the identifiers of one or more MTLFs associated with the analysis ID, and the MTLF in S101 may be one of the one or more MTLFs associated with the analysis ID.

[0138] The analysis identifier can be used to indicate an analysis service (or analysis service) associated with a model, i.e., the model can be used to perform the analysis service. The association of the MTLF with the analysis identifier can be understood as indicating that the model support provided by the MTLF is used to perform the analysis service corresponding to the analysis identifier.

[0139] Alternatively, whether the MTLF supports the analysis service corresponding to the analysis identifier may be determined by the NRF. For example, the NRF may be configured with an association between the MTLF and the identifiers of the analysis services supported by the MTLF. Therefore, as an optional implementation, the AnLF may send the analysis identifier to the NRF, and the NRF may provide the AnLF with the identifier of the MTLF corresponding to the analysis identifier.

[0140] In various embodiments of this application, the MTLF identifier may include at least one of the MTLF device identifier (MTLF NF ID), the MTLF vendor identifier (vendor ID), and the interoperability indicator. In this application, the device identifier may also be referred to as the NF Instance ID.

[0141] In this application, an MTLF NF ID may correspond to one MTLF. An MTLF or MTLF NF ID may correspond to one or more analysis identifiers. For example, MTLF NF ID 1 corresponds to analysis identifier 1 and analysis identifier 2, meaning that the MTLF to which MTLF NF ID 1 belongs corresponds to analysis identifier 1 and analysis identifier 2. Additionally, MTLF NF ID 2 may correspond to analysis identifier 1 and / or analysis identifier 2. In this application, the fact that an MTLF NF ID corresponds to an analysis identifier may mean that the MTLF can provide the analysis service corresponding to the analysis identifier.

[0142] The manufacturer identifier can be used to identify a device manufacturer. A manufacturer identifier can correspond to one or more MTLF device identifiers. For example, MTLF NF ID 1 and MTLF NF ID 2 can both correspond to manufacturer identifier 1, that is, the MTLF to which MTLF NF ID 1 belongs and the MTLF to which MTLF NF ID 2 belongs belong to the same manufacturer, and the manufacturer identifier of the manufacturer is manufacturer identifier 1. Optionally, the analysis identifier can be associated with the manufacturer identifier. The association of the analysis identifier with the manufacturer identifier can mean that the analysis identifier is associated with some or all of the manufacturer's MTLFs, that is, some or all of the manufacturer's MTLFs can provide the analysis services corresponding to the analysis identifier.

[0143] The interoperability identifier is a list of MTLF providers, for example, the interoperability identifier represents the manufacturer identifier, or the interoperability identifier is associated with the manufacturer identifier. The interoperability identifier can be associated with the analysis identifier, such as a one-to-one correspondence between the two, indicating that the MTLF allows the corresponding manufacturer or the MTLF included in the manufacturer to obtain the model corresponding to the analysis identifier, and / or indicates that the MTLF is allowed to interoperate with the AnLF on the model corresponding to the analysis identifier. Optionally, a MTLF may have one or more interoperability identifiers. If there are multiple interoperability identifiers, the multiple interoperability identifiers correspond to different analysis identifiers respectively. For example, MTLF NF ID 1 corresponds to analysis identifier 1 and analysis identifier 2, wherein the MTLF to which MTLF NF ID 1 belongs has interoperability identifier 1 and interoperability identifier 2, interoperability identifier 1 corresponds to analysis identifier 1, and interoperability identifier 2 corresponds to analysis identifier 2. Optionally, if the MTLF to which MTLF NF ID 1 belongs and the MTLF to which MTLF NF ID 2 belongs support interoperability, the MTLF to which MTLF NF ID 2 belongs may have interoperability identifier 1 and interoperability identifier 2, wherein interoperability identifier 1 corresponds to analysis identifier 1, and / or, interoperability identifier 2 corresponds to analysis identifier 2, that is, MTLFs of the same manufacturer may have the same interoperability identifier for the same analysis identifier. In addition, if the MTLF to which MTLF NF ID 1 belongs and the MTLF to which MTLF NF ID 2 belongs belong to different manufacturers, the MTLF to which MTLF NF ID 2 belongs may have interoperability identifier 3 and interoperability identifier 4, wherein interoperability identifier 3 corresponds to analysis identifier 1, and / or, interoperability identifier 4 corresponds to analysis identifier 2, that is, MTLFs of the same manufacturer may have different interoperability identifiers for the same analysis identifier.

[0144] In another implementation, MTLF1 identified by MTLF NF ID 1 belongs to vendor 1, and MTLF2 identified by MTLF NF ID 2 belongs to vendor 2. Optionally, if both MTLF1 and MTLF2 support interoperability for a certain analysis identifier (for example, analysis identifier 1), MTLF1 may have an interoperability identifier (vendor1, vendor2), and MTLF2 may have an interoperability identifier (vendor1, vendor2).

[0145] In an optional implementation of obtaining the identifier of the MTLF associated with the analysis identifier, the association relationship between the analysis identifier and the MTLF may be configured in the NRF. In this implementation, the AnLF can obtain the identifier of the MTLF associated with the analysis identifier from the NRF. Optionally, the AnLF can send the analysis identifier of the analysis service to be executed to the NRF. The NRF can learn, based on a pre-configuration, which MTLFs or manufacturers allow the AnLF to obtain the model associated with the analysis identifier provided by the AnLF, that is, which MTLFs or manufacturers support training models for executing the analysis service, and provide the identifiers of these MTLFs to the AnLF. This configuration may be provided by a management network element.

[0146] For example, as shown in Figure 4, the AnLF may provide the NRF with an analysis identifier, and the AnLF may obtain the device identifiers of one or more MTLFs associated with the analysis identifier from the NRF. Specifically, as shown in S201 and S202 in Figure 4, the AnLF may send an NF discovery (Nnrf_NFDiscovery) message to the NRF, which carries the analysis identifier, and the NRF may provide the AnLF with the identifiers of one or more MTLFs associated with the analysis identifier through a response message. As described above, the identifier of the MTLF may include at least one of a device identifier, a manufacturer identifier, and an interoperability identifier. Optionally, the response message may also carry an analysis identifier to indicate the association between the analysis service and the MTLF.

[0147] As an example of S101, the subscription message may include the index of the data to indicate the association relationship between the data and the model. For example, before S101, the index of the data has been generated by AnLF or MTLF or other network elements. When requesting the model through the subscription message, the index of the data is also carried to request the data. In addition, AnLF can also send the index of the data to MTLF through other messages. For example, before or after the subscription message, a message carrying the index of the data is sent to request the data. Optionally, the message carrying the index of the data may also carry a model identifier and / or an analysis identifier to indicate the association relationship between the model and the data. The model identifier can be provided at the same time when the MTLF provides the model to AnLF. For example, AnLF sends a subscription message to MTLF, and then MTLF sends the model and model identifier to AnLF, and then AnLF sends the index of the data and the model identifier to MTLF to indicate the association relationship between the model and the data.

[0148] In this application, data indexes can be used to indicate or obtain model-related data. For example, the data index can be a combination of an AnLF identifier and a dataset tag. In another example, the data index can be a dataset tag, in which case the dataset tag can include the AnLF identifier or a portion of the AnLF identifier. For example, the dataset tag can include the AnLF identifier or other information that uniquely identifies the NFp. The dataset tag can be used to identify a dataset that includes model-related data.

[0149] In this application, the identifier of AnLF may be the device identifier of AnLF, that is, AnLF NF ID.

[0150] For example, as shown in S203 in Figure 4, AnLF can send a model subscription message (MLModelProvision_Subscribe) to MTLF. This message is a subscription message, and S203 corresponds to S101. The model subscription message can be used by MTLF to obtain the model.

[0151] In addition, the model subscription message may further include an analysis identifier. Accordingly, the MTLF may carry the model and the model identifier for executing the corresponding analysis service in the response message shown in S204 according to the analysis identifier.

[0152] Optionally, the model subscription message may include a data index to establish an association between the requested model and the dataset or data. Alternatively, AnLF may also carry an analysis identifier in other messages to obtain the model from MTLF. For example, AnLF may request a model from MTLF via a subscription message, and then provide MTLF with the association between the model and the data index via a model subscription message or a model subscription update message, so that MTLF knows the index of the data related to the model and can request data based on the index.

[0153] Optionally, AnLF can indicate the index of data to ADRF so that ADRF stores the data related to the model to the index. For example, as shown in Figure 4, AnLF can send the resource storage request (Nadrf data management storage request) shown in S205 to ADRF to request ADRF to store data. Wherein, the request may include the index of data. For example, the index of data includes the combination of the identifier of AnLF and the data set index, then correspondingly, ADRF can adopt the combination of the identifier of AnLF and the data set index as the index of data when storing the model. Optionally, ADRF can send a storage transaction identifier (storage transaction identifier) ​​to AnLF via a response message, and the identifier can be the storage identifier that ADRF generates for the index of data. As shown in Figure 4, during the data storage process, ADRF can store the data related to the model derived from AnLF to the aforementioned index and mark it.

[0154] S102: AnLF sends an authorization message to ADRF, which is used to indicate that MTLF is allowed to obtain model-related data. The authorization message includes the MTLF's identifier and the index of the data. Correspondingly, ADRF stores the MTLF's identifier and the index of the data.

[0155] In this application, the authorization message may also be referred to as a data storage request message, a resource storage request message, etc.

[0156] The MTLF in S102 includes the MTLF in S101.

[0157] As an example, the authorization message shown in S102 may carry the identifier of the MTLF that is allowed to obtain the model-related data. Therefore, the ADRF can know which MTLFs are allowed to obtain the model-related data based on the identifier of the MTLF.

[0158] In addition, the index of the data may be used to indicate or determine the data that the MTLF is allowed to obtain or the data set where the data is located, that is, the index of the data is the index of the data that the MTLF is allowed to obtain.

[0159] As described in this application, the MTLF identifier may include at least one of a device identifier, a manufacturer identifier, or an interoperability identifier of the MTLF.

[0160] If the identifier of the MTLF is a device identifier of the MTLF, the MTLF indicated by the device identifier is allowed to obtain the data having the index.

[0161] Optionally, the authorization message may also include an analysis identifier, which may be associated with a manufacturer identifier and / or an interoperability identifier. If the identifier of the MTLF is a manufacturer identifier, then some or all of the MTLFs having the manufacturer identifier and corresponding to the analysis identifier of the model are allowed to obtain data having the index. Therefore, the MTLF having the manufacturer identifier and supporting the analysis business of the analysis identifier can be determined by combining the analysis identifier with the manufacturer identifier. If the identifier of the MTLF is an interoperability identifier, then all of the MTLFs having the interoperability identifier and corresponding to the analysis identifier of the model are allowed to obtain data having the index. Therefore, the MTLF having the interoperability identifier and supporting the analysis business of the analysis identifier can be determined by combining the analysis identifier with the interoperability identifier.

[0162] It is understandable that AnLF does not have to provide the analysis identifier to ADRF, but only needs to let ADRF know the index of the data that MTLF is allowed to obtain.

[0163] In addition, the authorization message may also include an AnLF identifier to indicate the source of the data.

[0164] In a possible embodiment, the authorization message shown in S102 may carry an identification set, which may include the identification of one or more MTLFs that are allowed to obtain model-related data, to indicate the MTLFs that are allowed to obtain model-related data. Accordingly, the identification set may also include the identification of one or more MTLFs, wherein, if multiple MTLF identifications are included, the multiple MTLFs may correspond to the same or different models. In this application, the identification set may also be referred to as an identification list, specifically a list of allowed network function consumers for the data, a list of allowed consumers for the data, or a list of allowed data consumers for the data, etc.

[0165] Optionally, if an identification set is carried, the identification set may also include an index of the data. For example, the identification set includes multiple lines of information, and the correspondence between an MTLF identification and an index of data may be included in one line of information in the identification set.

[0166] As shown in Table 1, the identification set may include identifications of one or more MTLFs, so that the one or more MTLFs can obtain data of one or more models from the ADRF.

[0167] Table 1

[0168] It can be understood that in Table 1, the multiple items of the MTLF NF ID, the vendor identifier, and the interoperability identifier corresponding to the index of the same data are in an "and" relationship. For example, the index 5 of the data corresponds to the MTLF NF ID 4, the vendor identifier 2, and the interoperability identifier 2, which means that the data corresponding to the index 5 is allowed to be identified as MTLF NF ID 4, has the vendor identifier 2 and corresponds to the analysis identifier 3, or has the interoperability identifier 2 and corresponds to the analysis identifier 3. All MTLF acquisitions, wherein the vendor identifier of the MTLF corresponding to the MTLF NF ID 4 is not the analysis identifier 3, and / or the interoperability identifier of the MTLF corresponding to the MTLF NF ID 4 is not the interoperability identifier 2.

[0169] It should be understood that Table 1 is merely an exemplary implementation of the identifier set. In actual applications, only a portion of the information shown in Table 1 may be included. For example, the AnLF NF ID and / or analysis identifier may not be included in the identifier set. Furthermore, appropriate expansion or modification may be made based on Table 1.

[0170] Taking Table 1 as an example, if the identifier of the MTLF requesting data exists in the identifier set, it means that AnLF allows MTLF to obtain the corresponding data, and ADRF can provide the MTLF with data with the corresponding index in response to the MTLF's data request. In addition, the data request can carry the index of the data. If the index of the data carried in the data request is consistent with the index of the data allowed to be obtained by MTLF, ADRF can provide the MTLF with data with the corresponding index in response to the MTLF's data request.

[0171] In this application, a data request may also be referred to as a data subscription message, a data message, or a data acquisition message.

[0172] In one possible embodiment, the authorization message in S102 may be the resource storage request shown in S205 in FIG4 , that is, the AnLF may carry the MTLF identifier or an identifier set including the MTLF identifier in the resource storage request. The authorization message may also be other messages other than the resource storage request, which is not specifically limited in this application.

[0173] Optionally, the authorization message may also include an authentication indication, which is used to indicate that the MTLF that obtains the data needs the authorization of AnLF. The reason is that not all types of data need to verify the authorization of AnLF when requested by MTLF, so AnLF needs to indicate the index of the data related to the model, so that ADRF knows which data is model-related data, or so that ADRF knows which data needs to verify the authority of MTLF. Accordingly, ADRF needs to provide data to MTLF based on the authentication information, that is, ADRF provides model-related data to the MTLF authorized by AnLF. Optionally, the authentication information can be used to indicate that the data is related to the model, and ADRF can be configured to verify the authority of the MTLF that requests the data before sending the data related to the model. Taking the method shown in Figure 3 as an example, the process of verifying the authority of MTLF may include: determining whether the identification of MTLF is included in the authorization information, and the authorization information may refer to the identification or identification set of MTLF provided by AnLF through the authorization message.

[0174] In addition, the authorization information can also be understood as being used to instruct MTLF to authorize AnLF to store model-related data in ADRF. Optionally, the data has an index or data set index specified by MTLF. For example, MTLF can provide a specified index to ADRF, such as generating the index by MTLF, and MTLF can also indicate to ADRF the identifier of AnLF that allows data to be stored in the index. Optionally, when providing the model to AnLF, MTLF can send the index of the data to indicate the association relationship between the model and the index of the data. Accordingly, before storing the data with the specified index from AnLF, ADRF can verify whether AnLF has obtained authorization from MTLF, that is, whether the identifier of AnLF is included in the range of the identifier of AnLF provided by MTLF.

[0175] In addition, the authentication indication can also be sent separately from the authorization message, or, when the index of the data is a special value, ADRF may assume that the data has an authentication indication, that is, the MTLF that requests the data by default needs to obtain authorization from AnLF, so the action of AnLF sending the authentication indication to ADRF can be omitted. In addition, it can also be considered that when there is a correspondence between the MTLF identifier and the index of the data in the identifier list or identifier set shown in Table 1, the MTLF needs to obtain authorization when requesting data with the corresponding index, so the authentication indication may not exist.

[0176] Optionally, in this application, ADRF may also assume that any data requested by MTLF or any model-related data needs to be authorized by AnLF, and in this case, AnLF does not need to send an authentication indication.

[0177] Based on the process shown in Figure 3, MTLF can obtain data from ADRF according to the subscription message in S102. Optionally, as shown in Figure 3, MTLF can send a data request to ADRF to request model-related data. The data request may carry the index of the data and the identifier of MTLF. The index of the data may include a combination of the identifier of AnLF and the dataset tag; or, the index of the data may include the dataset tag, and the dataset tag may include the identifier of AnLF. As shown in Figure 4, the data request may be a data management retrieval request (Nadrf_DataManagement_RetrievalRequest) message, or a resource retrieval request message, which may carry the index of the data (or storage transaction identifier). Accordingly, in the case where MTLF is allowed to obtain data related to the model, if the identifier of the MTLF is included in the identifier set, ADRF can provide data to MTLF according to the data request, that is, send model-related data to MTLF.

[0178] Optionally, ADRF can also determine whether there is an authentication indication for the data based on the identifier of AnLF and / or the index of the data, that is, ADRF can identify whether the MTLF that obtains the data needs authorization from AnLF.

[0179] In addition, the ADRF may further query the identifier or identifier set of the MTLF based on the index of the data to determine whether the MTLF is allowed by the AnLF to obtain the data. For example, if the MTLF with MTLF NF ID 1 requests to obtain the data identified by index 1, then according to Table 1, MTLF NF ID 1 is allowed to obtain the data. Therefore, in S104, the ADRF may send the data to the MTLF.

[0180] Optionally, if the identifier of MTLF is a manufacturer identifier or an interoperability identifier, MTLF can obtain a token containing the manufacturer identifier or interoperability identifier in the process of requesting a token from NRF. For example, as shown in Figure 4, MTLF can send a token message to NRF and provide the manufacturer identifier or interoperability identifier in the message. Correspondingly, after the identity authentication of MTLF, NRF can send a token to MTLF, and the token can contain the manufacturer identifier or interoperability identifier. MTLF can carry the token in the data request, and accordingly, ADRF can parse out the manufacturer identifier or interoperability identifier contained in the token. Further, if the manufacturer identifier or interoperability identifier of MTLF is carried in the authorization message in S103, ADRF can provide data to MTLF. Optionally, the token request can carry an analysis identifier, and accordingly, the token can also carry an analysis identifier to indicate that MTLF supports the analysis service corresponding to the analysis identifier.

[0181] It will be appreciated that before the MTLF sends a data request, it can obtain an authentication token from the NRF, such as an access token (access_token_nwdaf). Having obtained this token, the MTLF can act as an NFC and obtain services provided by the NFp. It will be appreciated that the token here is not used to verify whether the MTLF is permitted to obtain data provided by the AnLF; it is only used to verify whether the MTLF is permitted to obtain services as an NFC, such as obtaining permission to obtain the Data Management Retrieval (Nadrf_DataManagement_Retrieval) service. For example, as shown in Figure 4, the MTLF can send a token request to the NRF, and the NRF can send the token to the MTLF. Accordingly, the token can be included in the Data Management Retrieval request message shown in Figure 4.

[0182] For example, let's take the process of an NFc acquiring a token as an example. In a service-based system architecture, when an NFc requests a service from an NFp, the NFp needs to perform an authorization check on the requested service. Specifically, it checks whether the NFc is authorized to use the requested service. Only if the authorization check passes will the NFp provide the corresponding service to the NFc. For example, the ADRF might provide data to the MTLF. In this service, the ADRF acts as the NFp and the MTLF as the NFc. To verify the NFp's authorization for the service requested by the NFc, an access token-based authorization verification scheme can be used. Before sending the token, the NRF performs an authorization check on the NFc. This may include verifying the NFc's identity and, if the identity verification is successful, verifying whether the NFc has permission to access the requested service. If the NFc has permission, a token #1 is generated. This ensures that the service obtained by the NFc using token #1 is the authorized service, preventing unauthorized use of the service.

[0183] In addition, the NRF's verification of the NFc's identity may include verifying whether parameters carried in the token#1 acquisition request (e.g., the NF type of the NFc) match the NFc's public key certificate or NF configuration information. If they match, authentication succeeds; otherwise, authentication fails. Specifically, for example, the NRF may obtain the corresponding public key certificate from the NFc, which contains information about the NFc. The NRF compares the parameters in the token#1 acquisition request (e.g., the NF type of the NFc) with the information contained in the public key certificate. If the information matches, authentication succeeds; otherwise, authentication fails. Alternatively, for example, the NRF obtains NF configuration information corresponding to the NF identifier stored locally in the NRF network element (i.e., the NF configuration information of the NFc) based on the NFc's identifier in the token#1 acquisition request. The NRF then compares the parameters in the token#1 acquisition request, such as the NF type of the NFc, with the NF type in the NF configuration information. If they match, authentication succeeds; otherwise, authentication fails.

[0184] Optionally, NRF can perform security protection on the generated token#1. The security protection here can include integrity protection. For example, NRF uses a shared key to generate a message authentication code (MAC) for token#1, and NRF can send the MAC and token#1 together to NFc. The MAC can be used to verify whether the information in token#1 has been tampered with; or, NRF uses a private key to sign the content of token#1, and NFp can verify whether the content of token#1 has been tampered with through the signature. The specific method of using signatures and MACs to protect access tokens is defined in Request for Comments (RFC) 7515.

[0185] For example, if all parameters in the access token request message #1 are verified and the NFc is authorized to access the requested service, the NRF sends the generated token #1 to the NFc via an access token response message (e.g., Nnrf_AccessToken_Get Response). Optionally, the access token response message may also include other parameters, such as the expiration time (EXP) of token #1 or the scope of the request allowed to use token #1. In addition, if the NFc's authorization verification fails, the NRF sends an error response or a rejection response to the NFc.

[0186] It should be noted that the NRF verification in step S202 is performed at the service level. Therefore, if verification is successful, the NFc can initiate a service request based on token #1 to any NFp that can provide the service. Because the token verification is performed at the service level, obtaining the token does not necessarily grant the MTLF access to model-related data; it only means that the MTLF can obtain data services from the ADRF.

[0187] As can be seen, based on the processes shown in Figures 3 and / or 4, the ADRF can determine that the identifier of the MTLF requesting the model data is included in the authorization message (or resource storage request), so the MTLF allowed by the ADRF can obtain the data related to the model. As shown in Table 1, the identifier set may include the device identifier, manufacturer identifier, or interoperability identifier of the MTLF.

[0188] It can be understood that in the process shown in Figure 3 or Figure 4, the model-related data can be generated by the model, and / or, the model-related data can be used for training to generate the model.

[0189] Among them, when the data related to the model is generated by the model, AnLF can request the model from MTLF through the subscription message shown in S101 or the model subscription message shown in S203, thereby obtaining the model provided by MTLF. Further, based on the model, AnLF can obtain the data related to the model, such as the data related to the model includes model output data or model true value data. In addition, AnLF can also send the data related to the model to ADRF, and ADRF stores the data. In addition, AnLF can also indicate to ADRF that the MTLF is allowed to obtain the data related to the model through the authorization message shown in S102 or the resource storage request shown in S205. Therefore, MTLF can obtain the data related to the model from ADRF and determine whether to update the model based on the data related to the model.

[0190] When model-related data is used for training to generate the model, AnLF can generate and / or collect model-related data from other network elements, and send the model-related data to ADRF. At this time, the model-related data may include the input data of the model or the model's true value data, etc. Further, AnLF can also indicate to ADRF through the authorization message shown in S102 or the resource storage request shown in S205 that the MTLF is allowed to obtain the model-related data. Therefore, the MTLF can obtain the model-related data from ADRF and perform training based on the model-related data to obtain the model. In addition, AnLF can request the model from MTLF through the subscription message shown in S101 or the model subscription message shown in S203, thereby obtaining the model.

[0191] Different from the embodiments shown in Figures 3 and 4, as another embodiment, the AnLF can send a first token to the MTLF that provides the model, so that the MTLF requests model-related data from the ADRF based on the first token, and the ADRF sends the model-related data to the MTLF with the first token. Since the AnLF will not send the first token to the MTLF that has not provided the model, the ADRF will not send model-related data to the MTLF that has not provided the model, thereby improving the security of obtaining model-related data. As shown in Figure 5, an implementation method of this embodiment may include the following steps:

[0192] S301: AnLF obtains the identifier of MTLF.

[0193] The implementation of S301 may refer to the description of the method in which the AnLF obtains the identifier of the MTLF in S101, S201 or S202, and will not be repeated. For example, in S301, the AnLF may obtain the identifiers of one or more MTLFs from the NRF and determine a MTLF.

[0194] The MTLF identifier in S301 may be associated with the analysis identifier.

[0195] It can be understood that the MTLF identifier may include one or more of a MTLF device identifier, a manufacturer identifier, and an interoperability identifier.

[0196] S302: AnLF sends a subscription message including a first token to MTLF, where the subscription message is used to request a model.

[0197] In S302, the subscription information can refer to the description of the subscription message in S101, except that the subscription message in S302 additionally includes a first token for authorizing MTLF to obtain data of the model, that is, allowing MTLF to obtain data related to the model.

[0198] The first token may include an AnLF identifier, an MTLF identifier, and a data index. The AnLF identifier may be an AnLF NF ID. The MTLF identifier may include an MTLF device identifier, a manufacturer identifier, or an interoperability identifier. The data index may be used to obtain model-related data, and may be, for example, a combination of an AnLF identifier and a dataset index, or a dataset index including the AnLF identifier.

[0199] The first token can be protected by AnLF and / or NRF to prevent the first token from being tampered with. Security protection, such as integrity protection, can be implemented by referring to the scheme of NRF protecting tokens in the prior art.

[0200] Optionally, the first token may also include validity period information, such as an expiration date. AnLF may update MTLF's first token after the EXP expires. Furthermore, MTLF may request AnLF to update the first token after the EXP of the issued first token expires, using a machine learning model provision subscription notification (MLModelProvision_SubscribeNotify) message or a dataset index authorization request message.

[0201] Optionally, the first token may also include an analysis identifier, and the model is used to execute the analysis service corresponding to the analysis identifier. The analysis identifier can be used to indicate the MTLF together with the manufacturer identifier and / or interoperability identifier, that is, to determine the MTLF that has the manufacturer identifier and / or interoperability identifier and supports the analysis service corresponding to the analysis identifier.

[0202] As an example, the first token can be called a data token. The first token can be represented as token(NFp ID, NFc ID, datasetTag, EXP). The NFp ID represents the identifier of the MTLF. The NFc ID represents the identifier of the AnLF. The datasetTag represents the dataset index. The datasetTag can also be replaced with NFp ID+datasetTag, representing a combination of the AnLF identifier and the dataset index. Alternatively, the index of the data consisting of the combination of the AnLF identifier and the dataset index can be determined based on the NFp ID and datasetTag in the first token. EXP represents the expiration time.

[0203] In this embodiment, the first token is generated in a manner similar to the NRF's method for generating an authentication token, with the difference being that the first token is issued by the AnLF, while the authentication token is issued by the NRF. Only the MTLF permitted by the AnLF to obtain the model-related data, i.e., the MTLF that provides the model to the AnLF, can obtain the first token issued by the AnLF.

[0204] It is understood that in the embodiment shown in Figure 5, AnLF can issue a first token to MTLF based on its own needs and send the first token to MTLF via a subscription message. The process of issuing the token and sending the subscription message can be performed without the request of MTLF.

[0205] The subscription message may also be used to instruct the MTLF to include the first token in the data request when requesting data with the index from the ADRF. For example, the MTLF may be pre-configured through control signaling or predefined methods so that upon receiving the subscription message carrying the first token, the MTLF is informed that the first token needs to be included in the data request.

[0206] Taking Figure 6 as an example, in S401, the AnLF can send a model subscription message to the MTLF, and the message carries the first token. This model subscription message is an example of the subscription message in S302. Correspondingly, the AnLF can receive a response message from the MTLF, such as one for confirming the subscription or carrying the model.

[0207] As an example of S302, the first token can also be sent independently of the subscription message. For example, in addition to sending the subscription message, the AnLF can also send the first token to the MTLF via a model subscription message or a model subscription update message. For example, the AnLF can request a model from the MTLF via a subscription message, and then provide the MTLF with the association between the model and the data index, as well as the first token, via a model subscription message or a model subscription update message.

[0208] S303: The MTLF sends a data request including a first token to the ADRF, requesting model-related data.

[0209] The data request may also carry the data index and the MTLF identifier.

[0210] Optionally, the data request may also include an AnLF identifier to indicate the producer of the requested data. The data index may be a combination of the AnLF identifier and the dataset index, or the data index may be the dataset index.

[0211] For example, as shown in FIG6 , MTLF may send a data management acquisition request message to ADRF, which may carry the first token and the index of the data (or storage transaction identifier), and optionally, may also include the identifier of MTLF and / or the identifier of AnLT.

[0212] S304: ADRF sends model-related data to MTLF.

[0213] Among them, ADRF can send the data according to the first token, for example, when it is determined according to the first token that MTLF is allowed to obtain the data, the data is sent. The situation in which it is determined according to the first token that MTLF is allowed to obtain the data may include: ADRF verifies that the index of the data in the data request matches the index of the data in the first token. In addition, the data request may also include the identifier of AnLF, and the situation in which it is determined according to the first token that MTLF is allowed to obtain the data may also include: ADRF verifies that the identifier of AnLF in the data request matches the identifier of AnLF in the first token.

[0214] Optionally, before S304, AnLF may also send an authorization message to ADRF, which includes at least one of the identifier of AnLF, the index of the data, and an authentication indication. The authentication indication is used to indicate that authorization is required to obtain the data identified by the index. Accordingly, ADRF can verify the first token of MTLF before sending the data with the index. The index of the data and / or the identifier of AnLF can be used by ADRF to verify whether the first token provides MTLF with access rights to the requested data. For example, if the first token contains the index of the data and the identifier of AnLF, ADRF can send the data to MTLF.

[0215] Still taking Figure 6 as an example, in S402, AnLF can send a resource storage request to ADRF, and carry the AnLF identifier, the data index and the authentication indication in the message. The resource storage request can be used as an example of an authorization message. Accordingly, ADRF can use the index of the data or the storage transaction identifier to store data related to the model, and store the association relationship between the AnLF identifier, the data index and the authentication indication, that is, it is known that MTLF needs authorization when obtaining the data of the AnLF. In addition, ADRF can send the storage transaction identifier to AnLF via a response message. Among them, S402 and the sending of the response message can also be executed before S401.

[0216] Accordingly, before S304, ADRF can determine whether there is a corresponding authentication indication based on the index of the data in the data request and / or the identifier of AnLF. If so, that is, the authorization information carrying the index of the data and the authentication indication sent by AnLF has been received, then the first token of MTLF is verified.

[0217] Optionally, the authorization message may also include an analysis identifier, which is associated with the manufacturer identifier and / or interoperability identifier. Based on the authorization message, the ADRF can ensure that the MTLF that has the manufacturer identifier and / or interoperability identifier and supports the analysis service corresponding to the analysis identifier obtains data. Accordingly, the first token may also include an analysis identifier.

[0218] Exemplarily, ADRF may verify whether the index of the data contained in the first token is consistent with the index of the data contained in the authorization message, that is, MTLF is not allowed to obtain data other than the data authorized by AnLF, and / or verify whether the index of the data contained in the first token is consistent with the index of the data contained in the data request. In addition, ADRF may also verify whether the identifier of AnLF contained in the first token is consistent with the identifier of the producer AnLF in the authorization message, and / or verify whether the identifier of MTLF contained in the first token is consistent with the identifier of the MTLF that sent the data request. In addition, ADRF may also verify whether the EXP in the first token has expired.

[0219] For example, in the process shown in Figure 6, after the ADRF determines, based on the first token, that the MTLF has permission to obtain the data corresponding to the index, it may execute S404, i.e., send the data to the MTLF. It will be understood that steps not separately described in Figure 6 can be referred to the description of the response steps in the process of Figure 4 and will not be repeated here.

[0220] Optionally, if MTLF does not include the first token in the data request when requesting data from ADRF, and there is a corresponding authentication indication based on the data, ADRF can decide whether to provide the corresponding data to MTLF based on the local configuration policy, or decide to reject the request. For example, if MTLF does not have the authority, ADRF can optionally verify through signaling that AnLF has the data authority of AnLF, for example, by sending the MTLF's identifier and the index of the data requested by MTLF to AnLF. Alternatively, when the first token is not provided, ADRF can directly reject MTLF's request.

[0221] 5 and / or 6 , AnLF can send a first token to the MTLF that provides the model, indicating that the MTLF is allowed to obtain data related to the model, thereby preventing ADRF from providing data related to the model to the MTLF that does not have the first token.

[0222] It can be understood that in the process shown in Figure 5 or Figure 6, the model-related data can be generated by the model, and / or, the model-related data can be used for training to generate the model.

[0223] Among them, when the data related to the model is generated by the model, AnLF can request the model from MTLF through the subscription message shown in S302 or the model subscription message shown in S401, thereby obtaining the model provided by MTLF. Further, based on the model, AnLF can obtain the data related to the model, such as the model-related data including the model output data or the model true value data. In addition, AnLF can also send the data related to the model to ADRF, which stores the data. In addition, MTLF can also request and obtain the data related to the model from ADRF through the data request containing the first token shown in S303 or the data management acquisition request message containing the first token shown in S403. Therefore, after obtaining the data related to the model from ADRF, MTLF can determine whether to update the model based on the data related to the model.

[0224] When model-related data is used for training to generate the model, AnLF can generate and / or collect model-related data from other network elements, and send the model-related data to ADRF. At this time, the model-related data may include the input data of the model or the model's true value data, etc. Furthermore, AnLF can also request and obtain the model-related data from ADRF through the data request containing the first token shown in S303 or the data management acquisition request message containing the first token shown in S403. Therefore, after obtaining the model-related data from ADRF, MTLF can train according to the model-related data to obtain the model. In addition, AnLF can request the model from MTLF through the subscription message shown in S302 or the model subscription message shown in S401, thereby obtaining the model.

[0225] As another embodiment, the NRF can provide a second token to the MTLF that the AnLF allows to obtain data, so that the MTLF requests the AnLF's data from the ADRF based on the second token, and the ADRF sends the data to the MTLF with the second token. Optionally, the AnLF will not send the identifiers of other MTLFs that have not provided the model to the NRF, so the NRF will not send the second token to these MTLFs, and the ADRF will not send model-related data to these MTLFs, thereby improving the security of obtaining model-related data. As shown in Figure 5, an implementation method of this embodiment may include the following steps:

[0226] S501: The MTLF sends a token request to the NRF, including the AnLF identifier.

[0227] In response, the NRF receives the token request.

[0228] Optionally, before S501, the AnLF may send a subscription message to the NRF to request a model, so that the MTLF can obtain the identifier of the AnLF.

[0229] In addition, the MTLF may also be pre-configured with one or more identifiers of AnLFs that the MTLF can provide services for.

[0230] As shown in FIG8 , the token request may be included in the token message (Nnrf_token) shown in S603 , or in other words, the token message may include the token request.

[0231] S502: In the case where the MTLF is allowed to obtain the data provided by the AnLF, the NRF sends a second token to the MTLF, where the second token is used to authorize the MTLF to obtain the data.

[0232] Optionally, the NRF may also proactively send the second token to the MTLF while allowing the MTLF to obtain the data provided by the AnLF. That is, S501 may be optional.

[0233] Accordingly, the MTLF receives the second token.

[0234] Prior to S502, the AnLF may send the MTLF's identifier to the NRF to indicate that the MTLF is permitted to obtain the data provided by the AnLF. Optionally, the AnLF may also send the index of the data to the NRF to indicate that the MTLF is permitted to obtain the data with the index. For example, the MTLF's identifier is included in the identifier set sent by the AnLF to the NRF. The identifier set may also be referred to as a list, a list of permitted network function consumers for the data, a list of permitted consumers for the data, or a list of permitted data consumers for the data. Reference may be made to the description of the list shown in Table 1.

[0235] Accordingly, the NRF receives and / or stores the identifier of the MTLF.

[0236] The MTLF in S502 may be the MTLF that AnLF allows to obtain data provided by AnLF. Specifically, the data may be data related to the model provided by MTLF.

[0237] Exemplarily, the MTLF identifier may include at least one of a device identifier, a manufacturer identifier, and an interoperability identifier of the MTLF.

[0238] As an example of AnLF sending the MTLF identifier to NRF, as shown in FIG8 , AnLF can send the NF Management NF Registration (Nnrf_NFManagement_NFRegister) message shown in S601 to NRF, which includes the MTLF identifier, for example, carries an identifier set. Optionally, if the MTLF identifier includes a manufacturer identifier and / or an interoperability identifier, the message may also carry an analysis identifier for determining the MTLF that has the manufacturer identifier and / or interoperability identifier and supports the analysis service corresponding to the analysis identifier. Accordingly, NRF can send a response message to indicate that the message has been received.

[0239] It is understood that in S502, the MTLF is allowed to obtain the data provided by the AnLF, which may mean that the AnLF indicates the MTLF identifier to the NRF, indicating that the MTLF is allowed to obtain the AnLF data. For example, the identifier of the MTLF that sends the token request is included in the NF management NF registration message shown in S601, or is included in the identifier set in the NF management NF registration message.

[0240] If the MTLF is authenticated by the NRF, then in S604 , the NRF may send a second token to the MTLF.

[0241] Optionally, the AnLF may also send a subscription message to the MTLF, requesting a model. As described above, the MTLF may obtain the AnLF identifier based on this message. Alternatively, in a scenario where the MTLF obtains the AnLF identifier based on preconfiguration, the MTLF may receive the subscription message before or after sending the token request. This application does not specifically limit the execution sequence of the subscription message and the token request.

[0242] The subscription message may refer to the subscription message in S101. For example, the subscription message may include an index of the data. Optionally, the subscription message may also include an AnLF identifier.

[0243] It can be understood that after obtaining model-related data based on the process shown in Figure 5, MTLF can obtain the model based on data training and provide the model to AnLF in response to the subscription message, and / or, after providing the model to AnLF in response to the subscription request, it can obtain model-related data based on the process shown in Figure 5 and determine whether to update the model based on the data.

[0244] Taking Figure 6 as an example, in S602, AnLF can send a model subscription message to MTLF, which can be used as an example of a subscription message. Correspondingly, MTLF can send a response message to AnLF to indicate receiving the subscription or carrying the model.

[0245] In a possible embodiment, the second token in S502 includes an AnLF identifier and an MTLF identifier. The AnLF identifier may be an AnLF NF ID. The MTLF identifier may include an MTLF device identifier, a manufacturer identifier, or an interoperability identifier.

[0246] Optionally, the second token may also include an index of the data, an identifier of the ADRF, and / or an analysis identifier. The index of the data can be used to obtain model-related data to indicate which data the MTLF is allowed to obtain. The identifier of the ADRF can be used to determine the ADRF to indicate which ADRF the MTLF is authorized to obtain data from. The analysis identifier can be used together with the manufacturer identifier and / or interoperability identifier to determine the MTLF that has the manufacturer identifier and / or interoperability identifier and supports the analysis business corresponding to the analysis identifier.

[0247] Unlike the aforementioned authentication token, this second token is only issued to the MTLF indicated by the AnLF to the NRF. This MTLF can be used to provide the model to the AnLF, thereby preventing unauthorized MTLFs from obtaining the second token and using it to access AnLF data. Optionally, the second token may also include a data index, such as an index used to obtain data related to the model provided by the MTLF to the AnLF, further preventing unauthorized MTLFs from obtaining the model-related data.

[0248] It is understood that, compared to the verification process for NRF-issued authentication tokens in the prior art, the NRF's verification of the MTLF's second token requires additional verification to determine whether the MTLF is permitted to access the data provided by the AnLF. Furthermore, compared to the authentication token, the second token additionally includes the AnLF identifier and, optionally, an analysis identifier. Beyond this, the verification and generation of the second token can be similar to that of the authentication token.

[0249] S503: The MTLF sends a data request including the second token to the ADRF.

[0250] The data request may be used to request data provided by AnLF, such as model-related data.

[0251] In addition, the data request may also include a data index for identifying the data provided by AnLF.

[0252] Optionally, the data request may further include an AnLF identifier to indicate the producer of the requested data.

[0253] In this embodiment, the MTLF may optionally determine whether it has obtained the second token before requesting data from the ADRF based on the subscription message. If the second token has not been obtained, the MTLF may send the aforementioned token request to the NRF. Alternatively, if the token request sent by the MTLF to the NRF is rejected, it indicates that the MTLF is not authorized by the AnLF to obtain model-related data, and the NRF may stop requesting data from the ADRF.

[0254] The data request may include the data management acquisition request message shown in S605 in FIG8 .

[0255] S504: ADRF sends model-related data to MTLF according to the second token.

[0256] Among them, ADRF can send model-related data when MTLF is allowed to obtain data provided by AnLF. The situation in which MTLF is allowed to obtain data provided by AnLF includes, for example: the identifier of MTLF in the data request is consistent with the identifier of MTLF in the second token. In addition, the data request may also include the identifier of AnLF, and the situation in which MTLF is allowed to obtain data provided by AnLF may also include: the identifier of AnLF in the data request is consistent with the identifier of AnLF in the second token. In addition, the data request may also include the index of the data, and the situation in which MTLF is allowed to obtain data provided by AnLF may also include: the index of the data in the data request is consistent with the index of the data in the second token. In addition, the data request may also include the identifier of ADRF, and the situation in which MTLF is allowed to obtain data provided by AnLF may also include: ADRF determines that its own identifier is consistent with the identifier of ADRF in the second token.

[0257] Optionally, ADRF may also receive an authorization message from AnLF before S504. The authorization message may include an authentication indication, indicating that the MTLF that obtains the data needs authorization. Based on the authentication indication, ADRF can verify the MTLF's data acquisition authority after receiving the MTLF's authorization request.

[0258] In addition, the authorization message may also include at least one of an identifier of the MTLF, an index of the data, and an authentication indication. For the authorization message, please refer to the description in S304.

[0259] The authorization message can be seen in the resource storage request shown in FIG8 .

[0260] In S504, ADRF can verify whether the second token contains the identifier of AnLF, that is, verify whether AnLF is the data producer of the data requested by MTLF. In addition, ADRF can also verify whether the identifier of MTLF contained in the second token is consistent with the identifier of MTLF of the requested data, that is, verify whether the MTLF of the requested data is allowed by AnLF to obtain the data. Optionally, in the case that the second token also contains the index of the data, ADRF can also verify whether the index of the data contained in the second token is consistent with the index of the requested data, that is, verify whether the index of the data in the data request is consistent with the data index in the second token. Optionally, in the case that the second token also contains the identifier of ADRF, ADRF can also verify whether the identifier of ADRF contained in the second token is consistent with its own identifier. Optionally, if the identifier of MTLF is a manufacturer identifier and / or an interoperability identifier, it can be combined with the analysis identifier in the second token to determine the MTLF that has the manufacturer identifier and / or interoperability identifier and supports the analysis business corresponding to the analysis identifier, that is, the second token can also contain an analysis identifier.

[0261] It can be understood that the steps not separately described in FIG8 can refer to the description of the corresponding steps in the process of FIG4 and / or FIG6, and the description is not repeated.

[0262] It can be understood that in the process shown in Figure 7 or Figure 8, the model-related data can be generated by the model, and / or, the model-related data can be used for training to generate the model.

[0263] When model-related data is generated by the model, the AnLF can request the model from the MTLF via a subscription message in the embodiment of FIG. 7 or a model subscription message shown in S602, thereby obtaining the model provided by the MTLF. Furthermore, based on the model, the AnLF can obtain model-related data, such as model output data or model true value data. Furthermore, the AnLF can send the model-related data to the ADRF, which will store the data. Furthermore, the MTLF can request and obtain a second token from the NRF via a token request message containing the AnLF identifier shown in S501 or a token message containing the AnLF identifier shown in S603, thereby authorizing the MTLF to obtain the model-related data from the ADRF. The AnLF can send the MTLF identifier to the NRF to authorize the MTLF to obtain the data stored by the AnLF in the ADRF. The MTLF can further request and obtain the model-related data from the ADRF via a data request message containing the second token shown in S503 or a data management acquisition request message containing the second token shown in S605. After obtaining the model-related data from the ADRF, the MTLF can determine whether to update the model based on the model-related data.

[0264] When model-related data is used for training to generate the model, the AnLF can generate and / or collect model-related data from other network elements and send the model-related data to the ADRF. The model-related data may include model input data or model ground truth data. The MTLF can request and obtain a second token from the NRF via a token request message containing the AnLF identifier (S501) or a token message containing the AnLF identifier (S603). This authorizes the MTLF to obtain the model-related data from the ADRF. The AnLF can send the MTLF identifier to the NRF to authorize the MTLF to obtain the data stored by the AnLF in the ADRF. Furthermore, the MTLF can request and obtain the model-related data from the ADRF via a data request message containing the second token (S503) or a data management acquisition request message containing the second token (S605). After obtaining the model-related data from the ADRF, the MTLF can perform training based on the model-related data to obtain the model. Alternatively, the AnLF can request the model from the MTLF via a subscription message (S302) or a model subscription message (S401).

[0265] It can be seen that based on any embodiment shown in Figures 3 to 8 of the present application, on the one hand, MTLF needs to be authorized to obtain model-related data, which can improve security and prevent unauthorized MTLF or other network elements from obtaining data related to the model and further obtaining a model based on the data; on the other hand, MTLF has the authority to obtain the data from ADRF to train or update the model, thereby providing the trained model to AnLF, or providing an updated model.

[0266] It can be understood that in various embodiments of the present application, AnLF can authorize MTLF to obtain data related to the model when it hopes that a certain MTLF will continue to monitor the model, so that MTLF can monitor the accuracy of the model based on the relevant data, and / or provide a retrained model. Therefore, AnLF can determine one or more MTLFs associated with the analysis identifier, send a subscription message to the MTLF, and send an authorization message to the ADRF. In particular, the present application does not limit the sending sequence of subscription messages and authorization messages in any of the embodiments of Figures 3 to 8, nor does it limit the timing of model acquisition and sending subscription messages and / or authorization messages, because the sending of subscription messages and authorization messages can be for the index of data, so AnLF can first send subscription messages and / or authorization messages for the index of data, and then associate the index of data with a model, such as indicating the association between the model and the index of data to MTLF through the model acquisition process. AnLF can also first associate the index of data with a model, and then send subscription messages and / or authorization messages.

[0267] As an optional solution in this application, AnLF can also provide authorization for model-related data to other MTLFs other than the MTLF that provides the model, so that other MTLFs can test the accuracy of the model after obtaining the data, or use it to update the model.

[0268] It is understood that in various embodiments of the present application, when there are multiple MTLF identifiers associated with an analysis identifier, AnLF can select one MTLF from the multiple MTLFs based on its own decision and send a subscription message to the MTLF to request a model. In addition, AnLF can authorize the MTLF to obtain data related to the model. Therefore, even if there are other MTLFs that support the analysis identifier, if the model is not provided to AnLF, the MTLF cannot obtain authorization from AnLF and therefore cannot obtain data related to the model. It is understood that if AnLF needs multiple models, AnLF can request multiple models from one or more MTLFs, and accordingly, it can authorize one or more MTLFs to obtain data from the multiple models. The specific method can be referred to the description in this application and will not be repeated.

[0269] In each flow chart of the present application, dotted lines represent optional actions.

[0270] It is understandable that in order to implement the functions in the above embodiments, AnLF, MTLF, NRF and / or ADRF may respectively include hardware structures and / or software modules corresponding to the above functions. Those skilled in the art should easily realize that, in combination with the units and method steps of each example described in the embodiments disclosed in this application, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application scenario and design constraints of the technical solution.

[0271] Figures 9 and 10 are schematic diagrams of the structures of possible communication devices provided by the embodiments of the present application. These communication devices can be used to implement the functions of the analysis function network element, model training function network element, network storage function network element, or data storage network element in the above method embodiments, and thus can also achieve the beneficial effects possessed by the above method embodiments. Among them, the analysis function network element is, for example, AnLF, the model training function network element is, for example, MTLF, the network storage function network element is, for example, NRF, and the data storage network element is, for example, ADRF.

[0272] As shown in Figure 9, the communication device 900 includes a processing unit 910 and a transceiver unit 920. The communication device 900 is used to implement the functions of the analysis function network element, model training function network element, network storage function network element or data storage network element in any of the method embodiments shown in Figures 3 to 8 above.

[0273] When the communication device 900 is used to implement the function of the analysis function network element in the method embodiment shown in FIG3 : the transceiver unit 920 is used to send a subscription message and an authorization message.

[0274] When the communication device 900 is used to implement the function of the data storage network element in the method embodiment shown in FIG3 , the transceiver unit 920 may be configured to receive the authorization message and data request and send model-related data to the model training function network element. The processing unit 910 may be configured to determine whether the model training function network element is allowed to obtain the model-related data.

[0275] When the communication device 900 is used to implement the function of the analysis function network element in the method embodiment shown in Figure 5: the transceiver unit 920 is used to obtain the identifier of the model training function network element and send a subscription message containing the first token. The processing unit 910 can be used to generate the first token.

[0276] When the communication device 900 is used to implement the function of the data storage network element in the method embodiment shown in Figure 5: the transceiver unit 920 can be used to receive a data request containing a first token and send model-related data to the model training function network element.

[0277] When the communication device 900 is used to implement the function of the model training function network element in the method embodiment shown in Figure 5: the transceiver unit 920 is used to receive a subscription message containing the first token, send a data request containing the first token and receive model-related data.

[0278] When the communication device 900 is used to implement the function of the network storage function network element in the method embodiment shown in Figure 7: the transceiver unit 920 is used to receive the token request and send the second token. The processing unit 910 is used to obtain the second token.

[0279] When the communication device 900 is used to implement the function of the model training function network element in the method embodiment shown in Figure 7: the transceiver unit 920 is used to send a token request, receive a second token, send a data request containing the second token and receive model-related data.

[0280] For a more detailed description of the processing unit 910 and the transceiver unit 920, reference may be made to the relevant descriptions in any of the method embodiments shown in Figures 3 to 8. For example, the technical terms involved in the above actions of the processing unit 910 and the transceiver unit 920 may be referred to the descriptions in the above method embodiments and will not be repeated here.

[0281] As shown in Figure 10, communication device 1000 includes a processor 1010 and an interface circuit 1020. Processor 1010 and interface circuit 1020 are coupled to each other. It will be appreciated that interface circuit 1020 may be a transceiver or an input / output interface. Optionally, communication device 1000 may further include a memory 1030 for storing instructions executed by processor 1010, input data required by processor 1010 to execute instructions, or data generated by processor 1010 after executing instructions.

[0282] When the communication device 1000 is used to implement any of the methods shown in FIG. 3 to FIG. 8 , the processor 1010 is used to implement the functions of the processing unit 910 , and the interface circuit 1020 is used to implement the functions of the transceiver unit 920 .

[0283] When the above-mentioned communication device is a chip applied to an analysis function network element, a model training function network element, a network storage function network element or a data storage network element, the chip implements the functions of the analysis function network element, the model training function network element, the network storage function network element or the data storage network element in the above-mentioned method embodiment. The chip can receive information sent by other network elements or devices to the analysis function network element, the model training function network element, the network storage function network element or the data storage network element through other modules (such as radio frequency modules or antennas or communication interfaces) in the analysis function network element, the model training function network element, the network storage function network element or the data storage network element, or the chip sends information to other modules (such as radio frequency modules or antennas or interfaces) in the analysis function network element, the model training function network element, the network storage function network element or the data storage network element. Among them, the information is sent by the analysis function network element, the model training function network element, the network storage function network element or the data storage network element to other network elements or devices.

[0284] It is understood that the processor in the embodiments of the present application may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0285] The method steps in the embodiments of the present application can be implemented in hardware or in software instructions that can be executed by a processor. The software instructions can be composed of corresponding software modules, and the software modules can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disk, mobile hard disk, read-only memory (compact disc read-only memory, CD-ROM) or any other form of storage medium well known in the art. An exemplary storage medium is coupled to the processor so that the processor can read information from the storage medium and write information to the storage medium. The storage medium can also be an integral part of the processor. The processor and the storage medium can be located in an ASIC. In addition, the ASIC can be located in an analysis function network element, a model training function network element, a network storage function network element, or a data storage network element. The processor and the storage medium can also exist as discrete components in the analysis function network element, the model training function network element, the network storage function network element, or the data storage network element.

[0286] An embodiment of the present application also provides a communication system, including one or more network elements or devices selected from the group consisting of an analysis function network element, a model training function network element, a network storage function network element, and a data storage network element for implementing the above-mentioned method embodiment.

[0287] An embodiment of the present application further provides a computer-readable storage medium, which is used to store computer programs or instructions. When the computer-readable storage medium is executed, the method shown in the above method embodiment is implemented.

[0288] The embodiment of the present application also provides a computer program product, which, when executed on a computer, enables the method shown in the method embodiment to be implemented.

[0289] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are performed in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device, or other programmable device. The computer program or instructions may be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer program or instructions may be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a digital video disk; or a semiconductor medium, such as a solid-state drive. The computer-readable storage medium may be a volatile or nonvolatile storage medium, or may include both volatile and nonvolatile types of storage media.

[0290] In the description of this application, words such as "first" and "second" are used solely for descriptive purposes and should not be construed as indicating or implying relative importance or order. It should be noted that this application does not limit the order of appearance of "first," "second," etc. For example, "second" may appear before "first," and this is not a limitation in this application.

[0291] In the description of this application, "at least one (kind)" refers to one (kind) or more (kinds), and more (kinds) refers to two (kinds) or more than two (kinds). "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c can be single or multiple. In the description of this application, " / " means "or", for example, a / b means a or b.

[0292] In the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0293] It is understood that the various numbers used in the embodiments of this application are merely for ease of description and are not intended to limit the scope of the embodiments of this application. The order of the sequence numbers of the above-mentioned processes does not necessarily imply a specific order of execution; the order of execution of the processes should be determined by their functions and inherent logic.

Claims

1. A communication method, characterized in that: include: The analysis function network element sends a subscription message to the model training function network element, where the subscription message is used to request a model; The analysis function network element sends an authorization message to the data storage network element, where the authorization message is used to indicate that the model training function network element is allowed to obtain the data related to the model, and the authorization message includes the identifier of the model training function network element and the index of the data.

2. The method according to claim 1, characterized in that The authorization message includes the identifier of the model training function network element, including: The list of allowed network function consumers for the data in the authorization message includes the identifier of the model training function network element.

3. The method according to claim 1 or 2, characterized in that The index of the data is used to obtain data related to the model.

4. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: The analysis function network element sends the data to the data storage network element, and receives the model from the model training function network element, wherein the model is determined based on the data; or, The analysis function network element receives the model from the model training function network element, determines the data according to the model, and sends the data to the data storage network element.

5. The method according to claim 4, characterized in that The data is used to determine whether to update the model.

6. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: The analysis function network element sends an analysis identifier to the network storage function network element, and the model is used to execute the analysis service corresponding to the analysis identifier; The analysis function network element receives the identifier of the model training function network element from the network storage function network element, and the identifier of the model training function network element is determined based on the analysis identifier.

7. The method according to any one of claims 1 to 6, characterized in that: The subscription message includes an index of the data.

8. The method according to claim 1, characterized in that The identifier of the model training function network element includes one or more of the following: The device identifier of the model training function network element, the manufacturer identifier of the model training function network element, and the interoperability identifier of the model training function network element.

9. The method according to any one of claims 1 to 8, characterized in that: The authorization message also includes at least one of an identifier of the analysis function network element and an authentication indication, and the authentication indication is used to indicate that authorization is required to obtain the data.

10. A communication method, characterized in that: include: The data storage network element receives an authorization message, wherein the authorization message is used to indicate that the model training function network element is allowed to obtain model-related data, and the authorization message includes an identifier of the model training function network element and an index of the data; The data storage network element receives a data request from the model training function network element, where the data request is used to request the data; When the model training function network element is allowed to obtain the data related to the model, the data storage network element sends the data to the model training function network element.

11. The method according to claim 10, characterized in that The authorization message includes the identifier of the model training function network element, including: The list of allowed network function consumers for the data in the authorization message includes an identifier of the model training function network element; and in the case where the model training function network element is allowed to obtain the model-related data, the data storage network element sends the data to the model training function network element, comprising: In case the list of allowed network function consumers of the data includes the identifier of the model training function network element, the data storage network element sends the data to the model training function network element.

12. The method according to claim 11, characterized in that The method further comprises: The data storage network element stores the list of network function consumers allowed for the data.

13. The method according to any one of claims 10 to 12, characterized in that: The identifier of the model training function network element includes one or more of the following: The device identifier of the model training function network element, the manufacturer identifier of the model training function network element, and the interoperability identifier of the model training function network element.

14. The method according to any one of claims 10 to 13, characterized in that: The authorization message comes from the analysis function network element, and the method further includes: The data storage network element receives the data from the analysis function network element.

15. The method according to any one of claims 10 to 14, characterized in that: The data request includes the identifier of the analysis function network element and / or the index of the data.

16. The method according to any one of claims 10 to 15, characterized in that: The authorization message also includes at least one of an identifier of the analysis function network element and an authentication indication, and the authentication indication is used to indicate that authorization is required to obtain the data.

17. A communication method, characterized in that: include: Analyze the functional network element to obtain the identifier of the model training functional network element; The analysis function network element sends a subscription message to the model training function network element, where the subscription message is used to request a model. The subscription message includes a first token, where the first token is used to provide authorization for the model training function network element to obtain data related to the model. The first token includes an identifier of the analysis function network element, an identifier of the model training function network element, and an index of the data.

18. The method according to claim 17, characterized in that The index of the model-related data is used to obtain the model-related data.

19. The method according to claim 17 or 18, characterized in that The method further comprises: The analysis function network element sends the data to the data storage network element, and receives the model from the model training function network element, wherein the model is determined based on the data; or, The analysis function network element receives the model from the model training function network element, determines the data according to the model, and sends the data to the data storage network element.

20. The method of claim 19, wherein: The data is used to determine whether to update or train the model.

21. The method according to any one of claims 17 to 20, characterized in that: The first token further includes validity period information of the first token.

22. The method according to any one of claims 17 to 21, characterized in that: The method further comprises: The analysis function network element sends an authorization message to the data storage network element, wherein the authorization message includes an identifier of the analysis function network element, an index of the data, and at least one of an authentication indication, wherein the authentication indication is used to indicate that authorization is required to obtain the data.

23. The method according to any one of claims 17 to 22, characterized in that: The method further comprises: The analysis function network element receives a token update request from the model training function network element, where the token update request is used to update the first token.

24. The method according to any one of claims 17 to 23, characterized in that: The method further comprises: The data storage network element receives a data request from the model training function network element, where the data request is used to request the data and includes the first token; The data storage network element sends the data to the model training function network element according to the first token.

25. The method of claim 24, wherein: The data request includes an index of the data, The data storage network element sends the data to the model training function network element according to the first token, including: Upon verifying that the index of the data in the data request matches the index of the data in the first token, the data storage network element sends the data to the model training function network element.

26. The method according to claim 24 or 25, characterized in that The method further comprises: The data storage network element receives an authorization message, where the authorization message is used to indicate that authorization is required to obtain the data.

27. The method of claim 26, wherein: The authorization message includes one or more of an identifier of the analysis function network element, an index of the data, and an authentication indication, where the authentication indication is used to indicate that authorization is required to obtain the data.

28. The method according to claim 26 or 27, characterized in that The authorization message comes from the analysis function network element, and the method further includes: The data storage network element receives the data from the analysis function network element.

29. A communication method, characterized in that: include: The model training function network element receives a subscription message from the analysis function network element, the subscription message is used to request a model, the subscription message includes a first token, the first token is used to provide authorization for the model training function network element to obtain data related to the model, the first token includes an identifier of the analysis function network element, an identifier of the model training function network element, and an index of the data; The model training function network element sends a data request to the data storage network element, where the data request is used to request the data, and the data request includes the first token; The model training function network element receives the data from the data storage network element.

30. The method of claim 29, wherein: The index of the model-related data is used to obtain the model-related data.

31. The method according to claim 29 or 30, characterized in that The method further comprises: The model training function network element determines the model according to the data, and sends the model to the analysis function network element; or, The model training function network element sends the model to the analysis function network element, and the model is used to determine the data.

32. The method of claim 31, wherein: The model training function network element determines the model according to the data, and sends the model to the analysis function network element, including: the model training function network element determines to update the model according to the data, and sends the updated model to the analysis function network element.

33. The method according to any one of claims 29 to 32, characterized in that: The first token further includes validity period information of the first token.

34. The method according to any one of claims 29 to 33, characterized in that: The data request includes the identifier of the analysis function network element and / or the index of the data.

35. The method according to any one of claims 29 to 34, characterized in that: Also includes: The model training function network element sends a token update request to the analysis function network element, where the token update request is used to update the first token.

36. A communication method, characterized in that: include: The network storage function network element receives a token request from the model training function network element, wherein the token request includes an identifier of the analysis function network element; When the model training function network element is allowed to obtain the data provided by the analysis function network element, the network storage function network element sends a second token to the model training function network element, and the second token is used to provide authorization for the model training function network element to obtain the data. The second token includes the identifier of the analysis function network element and the identifier of the model training function network element.

37. The method of claim 36, wherein: The method further comprises: The network storage function network element receives a list of allowed network function consumers of the data from the analysis function network element, the list including an identification of the model training function network element, Among them, the situation that allows the model training function network element to obtain the data provided by the analysis function network element is that the list includes the identifier of the model training function network element.

38. The method according to claim 36 or 37, characterized in that The second token also includes an analysis identifier and / or an identifier of the data storage device, and the model is used to execute the analysis service corresponding to the analysis identifier.

39. The method according to any one of claims 36 to 38, characterized in that: The method further comprises: The network storage function network element receives the analysis identifier from the analysis function network element, and the model is used to execute the analysis service corresponding to the analysis identifier; The network storage function network element determines, according to the analysis identifier, the model training function network element that supports the analysis service corresponding to the analysis identifier; The network storage function network element sends the identifier of the model training function network element to the analysis function network element.

40. The method according to any one of claims 36 to 39, characterized in that: The method further comprises: The analysis function network element sends a list of allowed network function consumers for the data to the network storage function network element, and the list includes an identifier of the model training function network element.

41. The method according to any one of claims 36 to 40, characterized in that: The method further comprises: The analysis function network element sends the data to the data storage network element, and receives the model from the model training function network element, wherein the model is determined based on the data; or, The analysis function network element receives the model from the model training function network element, determines the data according to the model, and sends the data to the data storage network element.

42. The method of claim 41, wherein: The data is used to determine whether to update or train the model.

43. The method according to any one of claims 40 to 42, characterized in that: The method further comprises: The analysis function network element sends an analysis identifier to the network storage function network element, and the model is used to execute the analysis service corresponding to the analysis identifier; The analysis function network element receives the identifier of the model training function network element from the network storage function network element, and the identifier of the model training function network element is determined based on the analysis identifier.

44. The method according to any one of claims 36 to 43, characterized in that: The identifier of the model training function network element includes one or more of the following: The device identifier of the model training function network element, the manufacturer identifier of the model training function network element, and the interoperability identifier of the model training function network element.

45. A communication method, characterized in that: include: The model training function network element sends a token request to the network storage function network element, wherein the token request includes an identifier of the analysis function network element; The model training function network element receives a second token from the network storage function network element, where the second token includes an identifier of the analysis function network element and an identifier of the model training function network element; The model training function network element sends a data request to the data storage network element, where the data request is used to request the model-related data provided by the analysis function network element, and the data request includes the second token; The model training function network element receives the data from the data storage network element.

46. ​​The method of claim 45, wherein: The method further comprises: The model training function network element determines the model according to the data, and sends the model to the analysis function network element; or, The model training function network element sends the model to the analysis function network element, and the model is used to determine the data.

47. The method of claim 46, wherein: The model training function network element determines the model according to the data, and sends the model to the analysis function network element, including: the model training function network element determines to update the model according to the data, and sends the updated model to the analysis function network element.

48. The method according to any one of claims 45 to 47, characterized in that The second token also includes the analysis identifier and / or the identifier of the data storage device, and the model is used to execute the analysis business corresponding to the analysis identifier.

49. A communication device, characterized in that: The method comprises a processor configured to execute a computer program or instructions to implement the method according to any one of claims 1 to 48.

50. The device according to claim 49, characterized in that The device further comprises a memory and / or a transceiver, wherein the memory is used to store the computer program or instructions, and the transceiver is used for the device to communicate.

51. A computer program product, characterized in that The method comprises a computer program or instructions, which, when executed by a computer, causes the computer to implement the method according to any one of claims 1 to 48.

52. A communication device, characterized in that: Comprising units or modules for executing the method according to any one of claims 1-48.

53. A computer-readable storage medium, characterized in that: The storage medium stores a computer program or instructions. When the computer program or instructions are executed by the communication device, the method as described in any one of claims 1 to 48 is implemented.

54. A communication system, characterized in that: include: An analysis function network element, used to implement the method according to any one of claims 1 to 9; A data storage network element, used to implement the method according to any one of claims 10-16.

55. A communication system, characterized in that: include: An analysis function network element, used to implement the method according to any one of claims 17 to 23; A data storage network element, configured to execute the method according to any one of claims 24 to 28; A model training functional network element, used to implement the method as described in any one of claims 29-35.

56. A communication system, characterized in that: include: A network storage function network element, used to implement the method described in any one of claims 36-39 and 44; An analysis function network element, used to implement the method according to any one of claims 40 to 44; A model training functional network element is used to implement the method described in any one of claims 45-48.

Images