Service resource value determination method and device and access network equipment

Abstract

The invention relates to a service resource value determination method and device and access network equipment. The method comprises the following steps: under the condition that a service request is received, determining a service type corresponding to the service request; determining a target resource consumed in response to the service request according to the service type; the target resources comprise flow resources and computing power resources; obtaining a service resource value of the service request according to the target resource; the service resource value comprises a fusion resource value of the flow resource and the computing power resource. By adopting the method, combined charging of flow and computing power can be realized, and the complexity of charging management is reduced.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and in particular to a method, apparatus and access network equipment for determining service resource values. Background Technology

[0002] Currently, billing for user communication traffic in the network is mainly handled by the core network, while billing for computing resources relies primarily on an independent management system deployed on the edge cloud platform. The core network identifies and classifies incoming user traffic, determines service quality and service level based on preset policies, and uploads the traffic information to the billing system. This process typically only begins after user traffic traverses the radio access network and enters the core network, resulting in long billing response times. While edge cloud platforms can provide low-latency, high-performance computing services, their separation from the communication system and the lack of a unified metering dimension complicates billing management.

[0003] Therefore, current wireless communication technologies suffer from the problem of separate billing for data traffic and computing power, leading to complex billing management. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, access network equipment, computer-readable storage medium, and computer program product that can reduce the complexity of determining business resource values ​​in order to address the above-mentioned technical problems.

[0005] Firstly, this application provides a method for determining service resource values, applied to a wireless access network, including:

[0006] Upon receiving a business request, determine the business type corresponding to the business request;

[0007] Based on the business type, determine the target resources consumed in responding to the business request; the target resources include traffic resources and computing resources.

[0008] Based on the target resource, the service resource value of the service request is obtained; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

[0009] In one embodiment, determining the service type corresponding to the service request includes:

[0010] The service request is identified to obtain the service type; the service type includes at least one of communication type, computing type, and integrated computing type.

[0011] In one embodiment, determining the target resources consumed in responding to the service request based on the service type includes:

[0012] If the service type is the communication type, determine the traffic resources consumed in responding to the service request; or,

[0013] If the service type is the computing type, determine the computing resources consumed in responding to the service request; or,

[0014] When the service type is the integrated computing type, determine the traffic resources and computing resources consumed in responding to the service request.

[0015] In one embodiment, obtaining the service resource value of the service request based on the target resource includes:

[0016] When the service type is the communication type, the service resource value is obtained based on the traffic resources; or,

[0017] When the service type is the computing type, the service resource value is obtained based on the computing power resources; or,

[0018] When the service type is the integrated computing type, the traffic resources and the computing power resources are integrated to obtain the integrated resource value, and the service resource value is determined based on the integrated resource value.

[0019] In one embodiment, the step of fusing the traffic resources and the computing power resources to obtain the fused resource value includes:

[0020] The traffic resources are input into a first preset model to obtain traffic resource values; the first preset model reflects the mapping relationship between the traffic resources and the traffic resource values.

[0021] The computing resources are input into a second preset model to obtain computing resource values; the second preset model reflects the mapping relationship between the computing resources and the computing resource values.

[0022] The fused resource value is obtained by weighted summation of the traffic resource value and the computing power resource value.

[0023] In one embodiment, the computing power resources include the number of AI inferences and the amount of AI inference information. The second preset model includes a first sub-model and a second sub-model. The first sub-model reflects the mapping relationship between the number of AI inferences and the computing power resource value, and the second sub-model reflects the mapping relationship between the amount of AI inference information and the computing power resource value.

[0024] The step of inputting the computing resources into the second preset model to obtain computing resource values ​​includes:

[0025] Input the number of AI inference attempts into the first sub-model to obtain the first computing power resource value;

[0026] The AI ​​inference information is input into the second sub-model to obtain the second computing power resource value;

[0027] The first computing power resource value and the second computing power resource value are weighted and summed to obtain the computing power resource value.

[0028] In one embodiment, the method further includes:

[0029] According to a preset period, the business resource values ​​are reported to the core network and / or edge cloud platform.

[0030] Secondly, this application also provides a service resource value determination apparatus, applied in a wireless access network, comprising:

[0031] The type determination module is used to determine the business type corresponding to the business request upon receiving the business request.

[0032] The resource determination module is used to determine the target resources consumed in responding to the service request based on the service type; the target resources include traffic resources and computing power resources.

[0033] The numerical determination module is used to obtain the service resource value of the service request based on the target resource; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

[0034] Thirdly, this application also provides an access network device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0035] Upon receiving a business request, determine the business type corresponding to the business request;

[0036] Based on the business type, determine the target resources consumed in responding to the business request; the target resources include traffic resources and computing resources.

[0037] Based on the target resource, the service resource value of the service request is obtained; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

[0038] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0039] Upon receiving a business request, determine the business type corresponding to the business request;

[0040] Based on the business type, determine the target resources consumed in responding to the business request; the target resources include traffic resources and computing resources.

[0041] Based on the target resource, the service resource value of the service request is obtained; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

[0042] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0043] Upon receiving a business request, determine the business type corresponding to the business request;

[0044] Based on the business type, determine the target resources consumed in responding to the business request; the target resources include traffic resources and computing resources.

[0045] Based on the target resource, the service resource value of the service request is obtained; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

[0046] The aforementioned service resource value determination method, apparatus, access network equipment, computer-readable storage medium, and computer program product, upon receiving a service request, determine the service type corresponding to the service request, and based on the service type, determine the target resources consumed in responding to the service request. The target resources include traffic resources and computing power resources. Based on the target resources, the service resource value of the service request is obtained, which includes the combined resource value of traffic resources and computing power resources. The traffic resources and computing power resources consumed in responding to the service request can be determined on the radio access network side, and the final service resource value can be determined based on the traffic resources and computing power resources. Since the radio access network uniformly manages traffic billing and computing power billing, joint billing of traffic and computing power is realized, reducing the complexity of billing management. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is an application environment diagram of a method for determining business resource values ​​in one embodiment;

[0049] Figure 2 This is a flowchart illustrating a method for determining business resource values ​​in one embodiment;

[0050] Figure 3 This is a schematic diagram of the wireless intelligent management and orchestration function layer in one embodiment;

[0051] Figure 4 This is an interactive flowchart of a method for determining business resource values ​​in one embodiment;

[0052] Figure 5 This is a flowchart illustrating the method for determining business resource values ​​in another embodiment;

[0053] Figure 6 This is a structural block diagram of a service resource value determination device in one embodiment. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0055] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0056] The method for determining business resource values ​​provided in this application embodiment can be applied to, for example, Figure 1 In the application environment shown, terminal 102 communicates with the Radio Access Network (RAN) 104 via the network, and the RAN 104 transmits data with the Core Network (CN) 106 and the edge cloud platform 108 respectively.

[0057] Among them, the radio access network 104 can be a base station (NodeB, abbreviated as NB) with artificial intelligence (AI) function, an evolved base station (eNB or eNodeB), a next generation base station (gNB or gNodeB), etc., or an artificial intelligence radio access network (AI RAN) with a radio intelligent management and orchestration function layer (RAN AI Layer), which is not limited here.

[0058] Terminal 102 may be a wireless terminal, which can be a device providing voice and / or other service data connectivity to a user, a handheld device with wireless connectivity, or other processing devices connected to a wireless modem. The wireless terminal can communicate with one or more core networks via the RAN. The wireless terminal can be a mobile terminal, such as a mobile phone (or "cellular" phone) and a computer with a mobile terminal, for example, a portable, pocket-sized, handheld, computer-embedded, or vehicle-mounted mobile device, which exchanges voice and / or data with the wireless access network. The wireless terminal may also be referred to as a system, subscriber unit, subscriber station, mobile station, mobile, remote station, remote terminal, access terminal, user terminal, user agent, user device, or user equipment, without limitation herein.

[0059] The edge cloud platform 108 can be a cloud computing platform deployed on the edge, such as a multi-access edge computing (MEC) platform, a local dedicated edge cloud, etc., and is not limited here.

[0060] In traditional technologies, billing for user communication traffic in a network is primarily handled by the core network, while billing for computing resources mainly relies on independent management systems deployed on edge cloud platforms. Specifically, traffic billing depends heavily on the core network's policy and billing control modules. In 5G systems, the core network's User Plane Function (UPF) is responsible for identifying and classifying incoming user traffic, determining Quality of Service (QoS) and service level based on preset policies, and uploading traffic information to offline or online billing systems. This process typically begins only after user traffic traverses the RAN and enters the core network, resulting in long billing response times and unsuitability for low-latency scenarios. Furthermore, the core network billing system can only price communication traffic and cannot cover the computing power consumed by AI inference or model calls at the edge.

[0061] On the other hand, edge cloud platforms, as an emerging service delivery capability for operators, typically provide users with low-latency, high-performance computing services through MEC or local dedicated edge clouds. These platforms usually have independent computing power scheduling and billing systems, capable of tracking the consumption of Central Processing Unit (CPU) and Graphics Processing Unit (GPU) resources by users, and providing external application programming interfaces (APIs) for billing purposes. However, the separation between such computing platforms and communication systems results in users having to pay communication fees and service fees separately when using AI services, leading to complex billing practices.

[0062] Based on the aforementioned traditional technologies, this application provides a method for determining service resource values. Upon receiving a service request, the method determines the service type corresponding to the service request, and based on the service type, determines the target resources consumed in responding to the service request. The target resources include traffic resources and computing power resources. Based on the target resources, the method obtains the service resource value of the service request, which includes the combined resource value of traffic resources and computing power resources. The method can determine the traffic resources and computing power resources consumed in responding to the service request on the radio access network side, and determine the final service resource value based on the traffic resources and computing power resources. Since the radio access network uniformly manages traffic billing and computing power billing, it realizes joint billing of traffic and computing power, reducing the complexity of billing management.

[0063] It should be noted that the beneficial effects or technical problems solved by the embodiments of this application are not limited to this one, but may also be other implicit or related problems. For details, please refer to the description of the embodiments below.

[0064] In one exemplary embodiment, such as Figure 2 As shown, a method for determining business resource values ​​is provided. This embodiment applies this method to... Figure 1 Taking the radio access network 104 as an example, it can be understood that this method can be specifically applied to the RAN AI Layer of the radio access network 104. In this embodiment, the method includes the following steps:

[0065] Step S202: Upon receiving a business request, determine the business type corresponding to the business request.

[0066] Among them, a service request can be a message or instruction from a user requesting communication and / or computing services. For example, when a user requests a communication service, the service request can be a user access request; when a user requests a computing service, the service request can be a service initiation instruction. The service type can be the type to which the service request belongs, including but not limited to communication type, computing type, and converged computing type. Here, communication type refers to a user requesting communication services, computing type refers to a user requesting computing services, and converged computing type refers to a user requesting both communication and computing services.

[0067] Optionally, a user can send a service request to the AI ​​RAN. After receiving the service request, the RAN AI Layer can identify the service type to which the service request belongs. In some embodiments, the RAN AI Layer can intelligently identify and classify the received user access request (or service initiation instruction, etc.) to determine whether the user access request (or service initiation instruction, etc.) belongs to a communication type, a computing type, or a converged computing type.

[0068] Step S204: Determine the target resources consumed in responding to the business request based on the business type; the target resources include traffic resources and computing power resources.

[0069] The target resources can be the traffic resources and / or computing resources consumed by the RAN side in executing service requests. Traffic resources can be the communication traffic consumed in executing service requests, including but not limited to downlink and uplink communication traffic data. Computing resources can be the local computing resources invoked in executing service requests, including but not limited to GPU usage time and the number of times AI models are run.

[0070] For example, the RAN AI Layer can determine the traffic resources and / or computing resources consumed in responding to a service request based on different service types. In some embodiments, for communication-type service requests, the RAN AI Layer can collect downlink and uplink communication traffic data consumed by the RAN side in real time as target resources; for computing-type service requests, the RAN AI Layer can collect local computing resource information called for executing the service request in real time, such as GPU usage time and AI model execution times, as target resources; for converged computing-interconnected service requests, the RAN AI Layer can acquire downlink and uplink communication traffic data in real time as traffic resources, and simultaneously acquire local computing resource information as computing resources, using both traffic resources and computing resources as target resources.

[0071] Step S206: Based on the target resources, obtain the business resource value of the business request; the business resource value includes the combined resource value of traffic resources and computing power resources.

[0072] The business resource value can be the resource value consumed in responding to business requests. The converged resource value can be the total resource value of traffic resources and computing power resources. Resource value can be a metric value that converts resources into a unified unit for exchange, including but not limited to fees and value.

[0073] Optionally, the RAN AI Layer can determine corresponding resource values ​​for different target resources to obtain service resource values. In some embodiments, if the target resource is only traffic resource, the resource value of the traffic resource consumed by the service request can be calculated as the service resource value; if the target resource is only computing power resource, the resource value of the computing power resource consumed by the service request can be calculated as the service resource value; if the target resource is both traffic resource and computing power resource, the resource value of the traffic resource consumed by the service request and the resource value of the computing power resource consumed by the service request can be calculated, and the resource value of the traffic resource and the resource value of the computing power resource can be merged according to a preset weight, and the merged resource value can be used as the service resource value.

[0074] The aforementioned method for determining service resource values ​​involves, upon receiving a service request, determining the service type corresponding to the request, and based on the service type, determining the target resources consumed in responding to the request. These target resources include traffic resources and computing power resources. Based on the target resources, the service resource value for the request is obtained, which includes a combined value of traffic resources and computing power resources. The method allows the radio access network (RAN) to determine the traffic resources and computing power resources consumed in responding to the request, and then determine the final service resource value based on these resources. Since the RAN uniformly manages traffic billing and computing power billing, it achieves joint billing for traffic and computing power, reducing the complexity of billing management.

[0075] In an exemplary embodiment, step S202 may specifically include: identifying the service request to obtain the service type; the service type includes at least one of communication type, computing type, and converged computing type.

[0076] Among them, communication type refers to a user requesting communication services; computing type refers to a user requesting computing services; and integrated computing type refers to a user requesting both communication and computing services.

[0077] Optionally, after receiving a service request from a user, the RAN AI Layer can intelligently identify the service request based on AI algorithms to determine the service type to which the request belongs. For example, when a user requests communication services, the service type can be identified as communication; when a user requests computing services, the service type can be identified as computing; and when a user requests both communication and computing services, the service type can be identified as converged communication and computing.

[0078] It should be noted that the method for determining the business type is not limited to intelligent recognition. For example, business requests can also be classified based on deep learning, neural networks, artificial intelligence, and other methods to obtain the corresponding business type.

[0079] In this embodiment, by identifying the service request, the service type can be obtained quickly, which facilitates the adoption of different billing methods for different service types and improves billing efficiency.

[0080] In an exemplary embodiment, step S204 may specifically include: determining the traffic resources consumed in responding to a service request when the service type is communication; or determining the computing resources consumed in responding to a service request when the service type is computing; or determining the traffic resources and computing resources consumed in responding to a service request when the service type is integrated computing.

[0081] For example, if the RAN AI Layer determines that the service type is communication, it can obtain the traffic resources consumed in responding to the service request; if it determines that the service type is computing, it can obtain the computing power resources consumed in responding to the service request; if it determines that the service type is converged computing, it can obtain both the traffic resources and computing power resources consumed in responding to the service request. In some embodiments, for communication-type service requests, the downlink and uplink communication traffic data consumed by the RAN side can be collected in real time to obtain traffic resources; for computing-type service requests, the local computing power resource information called by the RAN can be statistically analyzed in real time, such as GPU usage time and AI model execution times, to obtain computing power resources; for converged computing-type service requests, downlink and uplink communication traffic data can be obtained in real time as traffic resources, and local computing power resource information can be obtained as computing power resources.

[0082] In this embodiment, by determining the traffic resources consumed in responding to a service request when the service type is communication, or the computing resources consumed in responding to a service request when the service type is computing, or the traffic resources and computing resources consumed in responding to a service request when the service type is integrated computing, billing can be performed separately for the resources consumed by different service types, ensuring the accuracy of billing.

[0083] In an exemplary embodiment, step S206 above may specifically include:

[0084] Step S301: If the service type is communication, obtain the service resource value based on the traffic resources; or,

[0085] Step S302: If the business type is computing, obtain the business resource value based on the computing power resources; or,

[0086] Step S303: When the business type is integrated computing type, the traffic resources and computing power resources are integrated to obtain the integrated resource value, and the business resource value is determined based on the integrated resource value.

[0087] For example, if the RAN AI Layer determines that the service type is communication, it can calculate the resource value consumed by the service request based on the determined traffic resources to obtain the service resource value; if the service type is computing, it can calculate the resource value consumed by the service request based on the determined computing power resources to obtain the service resource value; if the service type is converged computing, it can calculate the resource value of the traffic resources consumed by the service request based on the determined traffic resources, calculate the resource value of the computing power resources consumed by the service request based on the determined computing power resources, and merge the resource value of the traffic resources and the resource value of the computing power resources according to a preset weight to obtain a merged resource value, which is then used as the service resource value.

[0088] It is understandable that the process of merging the resource value of traffic resources and the resource value of computing power resources according to the preset weights can be, but is not limited to, a weighted summation of the resource values ​​of traffic resources and computing power resources.

[0089] In this embodiment, when the service type is communication, the service resource value is obtained based on traffic resources; when the service type is computing, the service resource value is obtained based on computing power resources; or when the service type is integrated computing, traffic resources and computing power resources are integrated to obtain an integrated resource value. The service resource value is determined based on the integrated resource value. This allows for separate billing of resources consumed by different service types, ensuring the accuracy of billing.

[0090] In an exemplary embodiment, step S303 above may specifically include:

[0091] Step S401: Input the traffic resources into the first preset model to obtain the traffic resource value; the first preset model reflects the mapping relationship between traffic resources and traffic resource value;

[0092] Step S402: Input the computing power resources into the second preset model to obtain the computing power resource value; the second preset model reflects the mapping relationship between computing power resources and computing power resource value;

[0093] Step S403: The traffic resource value and the computing power resource value are weighted and summed to obtain the fused resource value.

[0094] The first preset model can be a model that converts traffic resources into a metered value under a unified unit, such as a traffic billing model. The traffic resource value can be the resource value of traffic resources, such as traffic fees. The second preset model can be a model that converts computing power resources into a metered value under a unified unit, such as a computing power billing model. The computing power resource value can be the resource value of computing power resources, such as computing power fees. It should be noted that the billing model in this application embodiment can be, but is not limited to, a model set according to certain billing rules for calculating fees, representing the mapping relationship between consumed resources and fees.

[0095] Optionally, the RAN AI Layer can input traffic resources into a first preset model to obtain the traffic resource value output by the first preset model, input computing power resources into a second preset model to obtain the computing power resource value output by the second preset model, and then perform a weighted sum of the traffic resource value and the computing power resource value to obtain the fused resource value. For example, after acquiring uplink and downlink communication traffic data and the number of AI model inferences, the traffic cost corresponding to the uplink and downlink communication traffic data can be calculated according to a pre-known traffic billing model, and the computing power cost corresponding to the number of AI model inferences can be calculated according to a pre-known computing power billing model. The traffic cost and computing power cost can then be weighted and summed according to preset weights to obtain the total cost of traffic resources and computing power resources.

[0096] It is understood that communication-type or computing-type service requests can also use the method of this embodiment to calculate resource values. For example, for communication-type service requests, the weight of computing resource value can be set to 0, and the final fused resource value can be used as the service resource value, i.e., the communication cost; for computing-type service requests, the weight of communication resource value can be set to 0, and the final fused resource value can be used as the service resource value, i.e., the computing cost.

[0097] In this embodiment, traffic resources are input into a first preset model to obtain traffic resource values, and computing power resources are input into a second preset model to obtain computing power resource values. The traffic resource values ​​and computing power resource values ​​are weighted and summed to obtain a fused resource value. Traffic billing and computing power billing can be uniformly managed by the wireless access network, realizing joint billing of traffic and computing power and reducing the complexity of billing management.

[0098] In an exemplary embodiment, computing resources include the number of AI inference attempts and the amount of AI inference information. The second preset model includes a first sub-model and a second sub-model. The first sub-model reflects the mapping relationship between the number of AI inference attempts and the computing resource value, and the second sub-model reflects the mapping relationship between the amount of AI inference information and the computing resource value. Step S402 may specifically include: inputting the number of AI inference attempts into the first sub-model to obtain a first computing resource value; inputting the amount of AI inference information into the second sub-model to obtain a second computing resource value; and performing a weighted summation of the first computing resource value and the second computing resource value to obtain a computing resource value.

[0099] Here, the number of AI inference attempts can be the number of times the AI ​​model is invoked for inference. The amount of AI inference information can be the number of tokens input and / or output by the AI ​​model, where a token is the smallest unit of text processed by the model. The first sub-model can be a model that converts the number of AI inference attempts into a metric value under a unified unit, for example, a billing model for the number of AI inference attempts. The second sub-model can be a model that converts the amount of AI inference information into a metric value under a unified unit, for example, a billing model for the amount of AI inference information. The first computing power resource value can be the resource value for the number of AI inference attempts, for example, the cost of invoking the AI ​​model for inference n times. The second computing power resource value can be the resource value for the amount of AI inference information, for example, the cost of the AI ​​model inputting and outputting a total of m tokens.

[0100] For example, if the RAN AI Layer has counted the number of AI inferences required to execute a service request, it can input these counts into the first sub-model to obtain the first computing resource value corresponding to the number of AI inferences. If the RAN AI Layer has also counted the amount of AI inference information required to execute the service request, it can input this amount into the second sub-model to obtain the second computing resource value corresponding to the amount of AI inference information. The RAN AI Layer can then perform a weighted sum of the first and second computing resource values ​​to obtain the total computing resource value for executing the service request. For instance, when executing a computational service request, if the number of inferences n and the number of tokens m are recorded, the cost N for executing n inferences and the cost M for calling m tokens can be calculated based on a known billing model. The cost N and M can then be weighted and summed based on preset weights to obtain the total cost of computing resources consumed in executing the service request.

[0101] It is understandable that when only the number of inferences or the number of tokens called are counted, the method of this embodiment can also be used to determine the computing power resource value. For example, if only the number of inferences n is counted, the number of tokens called can be 0 or the weight of the number of tokens called can be 0, and the cost N of performing n inferences can be used as the total cost of computing power resources. If only the number of tokens called m is counted, the number of inferences can be 0 or the weight of the number of inferences can be 0, and the cost M of calling m tokens can be used as the total cost of computing power resources.

[0102] In this embodiment, the first computing power resource value is obtained by inputting the number of AI inferences into the first sub-model, and the second computing power resource value is obtained by inputting the amount of AI inference information into the second sub-model. The first computing power resource value and the second computing power resource value are weighted and summed to obtain the computing power resource value. The computing power resources consumed in executing business requests can be billed from the dimensions of the number of AI inferences and the amount of AI inference information, thereby improving the accuracy of billing.

[0103] In an exemplary embodiment, the above-described method for determining service resource values ​​may further include: reporting service resource values ​​to the core network and / or edge cloud platform according to a preset period.

[0104] The preset period can be a pre-set time period, such as a certain bill reporting period.

[0105] Optionally, after obtaining the service resource value of the service request based on the target resource, the RAN AI Layer can report the service resource value to the core network and / or edge cloud platform according to a preset period. In some embodiments, the RAN AI Layer can organize the service resource value into a data packet in a specified format (e.g., JSON or ASN.1 format) according to a certain billing reporting period and report it to the core network and / or edge cloud platform through a standardized interface.

[0106] In this embodiment, by reporting service resource values ​​to the core network and / or edge cloud platform according to a preset period, the cost of traffic resources and computing power resources can be determined on the wireless access network side, and the cost can be synchronized to the core network or edge cloud platform for the core network or edge cloud platform to perform relevant processing, thereby reducing service processing costs.

[0107] To facilitate a deeper understanding of the embodiments of this application by those skilled in the art, a specific example will be used for illustration below.

[0108] The existing communication-computing billing system is a separate architecture, with communication traffic billed by the core network and computing resources billed by the edge cloud platform, lacking a unified and integrated metering mechanism. In future AI-native business scenarios, such as smart glasses, humanoid guide dogs, and industrial machine vision, there is a greater demand for closer collaborative scheduling and package options for communication and computing integration. The existing technology system cannot reflect RAN-side resource consumption in real time, nor can it perform flexible and precise cost metering and policy execution at the access side, presenting several technical problems that urgently need to be addressed: First, as a centralized billing platform, the core network lacks real-time awareness of user-side communication resources at the RAN level. Its traffic billing relies on the uplink path, resulting in long links and high recognition latency, making it particularly unsuitable for low-latency services connecting to edge intelligent services. Second, the independent billing models for communication and computing resources prevent support for new package designs based on resource integration, and hinder effective billing support for business models such as pay-per-use AI services and inference duration billing. Third, with the development of AI RAN architecture, the RAN possesses native intelligent scheduling and inference execution capabilities, especially for future AI applications. General computing resources such as CPUs and GPUs deployed in the RAN have participated in user service processing, but there is currently a lack of mechanisms to record the usage of these resources, which in turn cannot be reflected in the billing. Ultimately, the RAN side has limited ability to understand the service context, and the current system has difficulty in integrating and identifying information such as model type, service intensity, and task level, making it difficult to support differentiated and dynamic pricing strategies.

[0109] In view of this, this application proposes a converged traffic and computing power billing method based on AI RAN. This method sinks the communication traffic billing function, originally located in the core network, and the computing power billing function, originally located in the edge cloud platform, to the RAN AI Layer. A converged billing mechanism is established on the AIRAN side, enabling unified monitoring and pricing of user communication and computing resources on the RAN side. The converged billing information is periodically reported to the core network and edge cloud platform, constructing a billing system that supports the integration and real-time billing of AI-native services. This system can support the construction of distributed packages and real-time billing capabilities for AI services. The method includes the following steps:

[0110] Step S501, Service Identification and Classification: After the RAN AI Layer receives a user access request or service initiation instruction, it intelligently identifies and classifies the service type, including communication type, computing type, and converged computing type services, etc.

[0111] Step S502, Collection of communication and computing resource usage: Real-time collection of downlink / uplink communication traffic data consumed by this service on the RAN side, as well as information on the local computing resources called up (such as GPU usage time, number of model runs, etc.).

[0112] Step S503, the billing management module performs converged pricing: the converged billing engine in the RAN AI Layer evaluates the combined consumption of communication and computing power according to the preset pricing strategy model and outputs a converged bill entry.

[0113] Among them, based on the deep context awareness capability of access services in the RAN AI Layer, the resource usage of traffic and computing power can be integrated. Through the preset integration strategy model (which can support AI inference model granularity, time granularity, dynamic weight, etc.), resource cost items with high accuracy, fine granularity and strong adaptability can be generated, realizing the real-time and fine-grained resource integration pricing capability that the traditional fragmented billing system cannot achieve.

[0114] Step S504: Periodic billing information is reported to the core network and edge cloud platform: The converged billing information is packaged into periodic billing data messages and reported to the core network and edge cloud through standardized interfaces, so that they can perform management functions such as package orchestration, policy distribution, and cross-domain settlement.

[0115] The aforementioned method effectively addresses key issues in existing wireless access networks, such as fragmented billing processes for communication and computing resources, high pricing response latency, and the inability to support customized packages for AI-native services. Specifically, this method achieves integrated resource pricing, unifying the perception and integrated pricing of communication traffic and computing power consumption through the RAN AI Layer, supporting multi-dimensional package designs based on communication and AI services. It also shortens the billing cycle and improves accuracy by generating billing entries in real-time on the RAN side, avoiding link latency caused by billing after service traffic is sent to the core network, thus improving billing response speed and accuracy. Furthermore, this method reduces system complexity and operating costs, eliminating the need to build a separate edge computing platform billing system for AI services, directly reusing the intelligent scheduling and context-aware mechanisms in the RAN AI Layer, reducing service deployment costs. Finally, this method supports real-time billing, package subscriptions, and billing reminders for AI services, enhancing service convenience and practicality. Therefore, the integrated billing technology adapted to AI RAN scenarios provided in this application can be widely applied to new business models such as smart terminal access, smart glasses services, embodied intelligent agents, and low-power AI devices.

[0116] In one exemplary embodiment, an AI RAN converged billing method is provided, applicable to end users using novel AI services such as embodied intelligence, and specifically includes the following steps:

[0117] Step S511, user initiates AI service request: terminal initiates AI navigation service request to base station through 5G access link, RAN AI Layer senses the service request and determines it to be a "communication + computing power" composite service through service management orchestration module;

[0118] Step S512, Real-time collection of resource usage: During service execution, the RAN AI Layer computing resource scheduling module collects the user's uplink and downlink traffic data, and at the same time records the local GPU resources called, including the number of model inferences, usage time, power consumption indicators, etc.

[0119] Step S513, generation of integrated billing entries: The triggering condition can be that the business processing exceeds the billing threshold or reaches the cycle node. The RAN AI Layer starts the billing management module, calls the pricing strategy model, and obtains the total cost of "communication + computing power".

[0120] The pricing strategy model can be set as: unit byte + unit calculation cycle × weight + token usage. The resource consumption is weighted and summarized, and a bill item can be generated. For example, communication 350MB + inference 10 times + 50,000 tokens, totaling 5.2 yuan.

[0121] Step S514, Bill Packaging and Reporting: Collect billing items into the current user cycle bill, organize them into data packets that conform to the core receiving format (supports JSON / ASN.1), and report them to the core network and edge cloud platform through the corresponding interface.

[0122] The aforementioned AI RAN converged billing method enables low-latency converged pricing, with the billing process independent of the core network. The RAN AI Layer completes traffic and computing power assessment locally, reducing billing latency. It also improves billing granularity and accuracy, generating billing entries at the user-service session level to achieve "pay-as-you-go" billing, adapting to the dynamic characteristics of AI services. Furthermore, it enhances network intelligence and operational flexibility, supporting operators in launching packages such as "X GB + X AI model calls + X Tokens per month," promoting the integration of communication and computing power.

[0123] Figure 3 A schematic diagram of a RAN AI Layer is provided. Figure 4 An interactive flowchart illustrating a method for determining business resource values ​​is provided. (Reference) Figure 3 and Figure 4 In one exemplary embodiment, a method for determining business resource values ​​is provided, specifically including the following steps:

[0124] Step S521: The terminal sends a service request to the AI ​​RAN;

[0125] In step S522, the service management and orchestration module of the RAN AI Layer identifies service requests and classifies them to obtain service types;

[0126] Step S523: The business management orchestration module requests communication resources from the general computing resource scheduling module or computing power resources from the model management service module according to the business type.

[0127] Step S524: The general computing resource scheduling module periodically reports the general computing resource usage status, and the model management service module periodically reports the model usage status.

[0128] In step S525, the billing management module generates a bill based on the communication resource usage and model usage, and reports it to the core network and edge cloud platform on a regular basis.

[0129] The aforementioned method for determining service resource values ​​adopts a measurement approach based on AI task granularity, rather than establishing a static binding relationship between users and physical computing resources. In AI RAN scenarios, user requests often trigger AI model calls in the form of intelligent tasks (such as path planning, network optimization, and image recognition). Therefore, the design uses tasks as the core resource unit. The RANAI Layer identifies tasks through the service orchestration module and coordinates with the general computing resource scheduling module to complete the required computing power call and resource recording. This approach adapts to the dynamic nature and task diversity of communication scenarios, offering greater flexibility and practical feasibility.

[0130] Secondly, regarding the integrated billing modeling of communication and computing power, a mechanism based on joint identification using user ID and task ID is proposed. This mechanism integrates and manages communication traffic data collected from the traditional RAN side with model usage data (such as type, frequency, duration, and resource consumption) collected from the AI ​​RAN side. This approach enables the statistical analysis and pricing of computing resources consumed by each AI task triggered by a user, providing a unified integrated billing path for multi-user, multi-task, and multi-resource sharing.

[0131] Finally, interfaces and mechanisms are reserved for subsequent multi-user resource isolation, differentiated services and refined billing strategies. The system is designed with collection mechanisms such as resource utilization, task execution statistics and model scheduling records, which provides basic support for operators to build a three-dimensional mapping relationship between users, tasks and resources. Dynamic authorization, priority scheduling and SLA guarantee functions can be realized in the future through policy configuration.

[0132] In summary, the above methods, through task-level scheduling, convergence identification mechanisms, and resource usage information collection, can effectively support the new converged billing requirements under the AI ​​RAN architecture.

[0133] In one exemplary embodiment, such as Figure 5As shown, a method for determining business resource values ​​is provided, including the following steps:

[0134] Step S601: Upon receiving a business request, identify the business request to obtain the business type;

[0135] Step S602: If the business type is a converged computing type, determine the traffic resources and computing power resources consumed in responding to the business request; the computing power resources include the number of AI inferences and the amount of AI inference information.

[0136] Step S603: Input the traffic resources into the first preset model to obtain the traffic resource value;

[0137] Step S604: Input the number of AI inferences into the first sub-model of the second preset model to obtain the first computing power resource value; input the amount of AI inference information into the second sub-model of the second preset model to obtain the second computing power resource value; and perform a weighted summation of the first computing power resource value and the second computing power resource value to obtain the computing power resource value.

[0138] Step S605: The traffic resource value and the computing power resource value are weighted and summed to obtain the fused resource value, and the business resource value is determined based on the fused resource value;

[0139] Step S606: According to the preset period, the business resource value is reported to the core network and / or edge cloud platform.

[0140] Optionally, a user can send a service request to the AI ​​RAN. The RAN AI Layer identifies the received service request and determines the corresponding service type. If the service type is a converged computing type, the RAN AI Layer can, in the process of responding to the service request, statistically analyze the consumed traffic and computing resources in real time. The computing resources can be the number of AI inference attempts and the amount of AI inference information. The RAN AI Layer can input the traffic resources into a traffic billing model to obtain a traffic resource value, input the number of AI inference attempts into a computing power billing model based on the number of inference attempts to obtain a first computing power resource value, and input the amount of AI inference information into a computing power billing model based on the number of tokens to obtain a second computing power resource value. The first and second computing power resource values ​​are then weighted and summed to obtain the total computing power resource value for the number of AI inference attempts and the amount of AI inference information. Finally, the traffic resource value and the computing power resource value are weighted and summed to obtain a converged resource value. This converged resource value is used as the service resource value consumed in responding to the user's service request. Since the specific processing procedure of the RAN AI Layer has been described in detail in the preceding embodiments, it will not be repeated here.

[0141] It is understandable that the weighted summation process of steps SS603 to S605 above can also be performed by weighted summation of the traffic resource value, the first computing power resource value and the second computing power resource value after calculating them, to obtain the business resource value. This will not be elaborated here.

[0142] The above-mentioned method for determining service resource values ​​can determine the traffic and computing resources consumed in responding to service requests on the radio access network side, and determine the final service resource value based on the traffic and computing resources. Since the radio access network uniformly manages traffic billing and computing power billing, it realizes joint billing of traffic and computing power, reducing the complexity of billing management.

[0143] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0144] Based on the same inventive concept, this application also provides a business resource value determination apparatus for implementing the business resource value determination method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more business resource value determination apparatus embodiments provided below can be found in the limitations of the business resource value determination method described above, and will not be repeated here.

[0145] In one exemplary embodiment, such as Figure 6 As shown, a business resource value determination device is provided, including: a type determination module 701, a resource determination module 702, and a value determination module 703, wherein:

[0146] The type determination module 701 is used to determine the service type corresponding to the service request upon receiving the service request.

[0147] The resource determination module 702 is used to determine the target resources consumed in responding to the service request based on the service type; the target resources include traffic resources and computing power resources.

[0148] The numerical determination module 703 is used to obtain the service resource value of the service request based on the target resource; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

[0149] In an exemplary embodiment, the type determination module 701 is further configured to identify the service request and obtain the service type; the service type includes at least one of communication type, computing type, and integrated computing type.

[0150] In an exemplary embodiment, the resource determination module 702 is further configured to determine the traffic resources consumed in responding to the service request when the service type is the communication type; or, when the service type is the computing type, determine the computing power resources consumed in responding to the service request; or, when the service type is the converged computing type, determine the traffic resources and computing power resources consumed in responding to the service request.

[0151] In an exemplary embodiment, the numerical determination module 703 is further configured to: obtain the service resource value based on the traffic resource when the service type is the communication type; or obtain the service resource value based on the computing power resource when the service type is the computing type; or, perform fusion processing on the traffic resource and the computing power resource to obtain the fused resource value when the service type is the converged computing type, and determine the service resource value based on the fused resource value.

[0152] In an exemplary embodiment, the numerical determination module 703 is further configured to input the traffic resource into a first preset model to obtain a traffic resource value; the first preset model reflects the mapping relationship between the traffic resource and the traffic resource value; input the computing power resource into a second preset model to obtain a computing power resource value; the second preset model reflects the mapping relationship between the computing power resource and the computing power resource value; and perform a weighted summation of the traffic resource value and the computing power resource value to obtain the fused resource value.

[0153] In an exemplary embodiment, the numerical determination module 703 is further configured to input the number of AI inferences into the first sub-model to obtain a first computing power resource value; input the amount of AI inference information into the second sub-model to obtain a second computing power resource value; and perform a weighted summation of the first computing power resource value and the second computing power resource value to obtain the computing power resource value.

[0154] In an exemplary embodiment, the above-mentioned service resource value determination device further includes a value reporting module, which is used to report the service resource value to the core network and / or edge cloud platform according to a preset period.

[0155] Each module in the aforementioned service resource value determination device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the access network device in hardware form or independent of it, or stored in the memory of the access network device in software form, so that the processor can call and execute the operations corresponding to each module.

[0156] In an exemplary embodiment, an access network device is provided, which can be an AI RAN. The access network device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor of the access network device provides computing and control capabilities. The memory of the access network device includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the access network device stores service resource value data. The I / O interfaces of the access network device are used for exchanging information between the processor and external devices. The communication interface of the access network device is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a service resource value determination method.

[0157] Those skilled in the art will understand that the above structure is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the access network equipment to which the present application is applied. Specific access network equipment may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0158] In one exemplary embodiment, an access network device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0159] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0160] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0161] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0162] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0163] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0164] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for determining the value of business resources, characterized in that, The method is applied to a wireless access network and includes: Upon receiving a business request, determine the business type corresponding to the business request; Based on the business type, determine the target resources consumed in responding to the business request; the target resources include traffic resources and computing power resources; Based on the target resource, the service resource value of the service request is obtained; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

2. The method according to claim 1, characterized in that, Determining the service type corresponding to the service request includes: The service request is identified to obtain the service type; the service type includes at least one of communication type, computing type, and integrated computing type.

3. The method according to claim 2, characterized in that, The step of determining the target resources consumed in responding to the service request based on the service type includes: If the service type is the communication type, determine the traffic resources consumed in responding to the service request; or, If the service type is the computing type, determine the computing resources consumed in responding to the service request; or, When the service type is the integrated computing type, determine the traffic resources and computing resources consumed in responding to the service request.

4. The method according to claim 2, characterized in that, The step of obtaining the service resource value of the service request based on the target resource includes: When the service type is the communication type, the service resource value is obtained based on the traffic resources; or, When the service type is the computing type, the service resource value is obtained based on the computing power resources; or, When the service type is the integrated computing type, the traffic resources and the computing power resources are integrated to obtain the integrated resource value, and the service resource value is determined based on the integrated resource value.

5. The method according to claim 4, characterized in that, The process of fusing the traffic resources and the computing power resources to obtain the fused resource value includes: The traffic resources are input into a first preset model to obtain traffic resource values; the first preset model reflects the mapping relationship between the traffic resources and the traffic resource values. The computing resources are input into a second preset model to obtain computing resource values; the second preset model reflects the mapping relationship between the computing resources and the computing resource values. The fused resource value is obtained by weighted summation of the traffic resource value and the computing power resource value.

6. The method according to claim 5, characterized in that, The computing power resources include the number of AI inferences and the amount of AI inference information. The second preset model includes a first sub-model and a second sub-model. The first sub-model reflects the mapping relationship between the number of AI inferences and the computing power resource value, and the second sub-model reflects the mapping relationship between the amount of AI inference information and the computing power resource value. The step of inputting the computing resources into the second preset model to obtain computing resource values ​​includes: Input the number of AI inference attempts into the first sub-model to obtain the first computing power resource value; The AI ​​inference information is input into the second sub-model to obtain the second computing power resource value; The first computing power resource value and the second computing power resource value are weighted and summed to obtain the computing power resource value.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: According to a preset period, the business resource values ​​are reported to the core network and / or edge cloud platform.

8. A device for determining the value of business resources, characterized in that, The device is used in a wireless access network and includes: The type determination module is used to determine the business type corresponding to the business request upon receiving the business request. The resource determination module is used to determine the target resources consumed in responding to the service request based on the service type; the target resources include traffic resources and computing power resources. The numerical determination module is used to obtain the service resource value of the service request based on the target resource; the service resource value includes the combined resource value of the traffic resource and the computing power resource.

9. An access network device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images