QoS assurance method, function, storage medium and computer program product

Abstract

The invention provides a quality of service (QoS) guarantee method and function, a storage medium and a computer program product, and the method comprises the steps that a first function receives first information sent by a second function, and the first information is used for describing a demand related to artificial intelligence (AI) service data transmission and / or a demand related to AI service computing power; generating a QoS control strategy of the AI service based on the first information and an AI service template corresponding to the signed AI service; and sending the QoS control strategy to the third function, wherein the QoS control strategy is used for the third function to generate a QoS parameter corresponding to the AI service based on the QoS control strategy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, and particularly relates to a Quality of Service (QoS) guarantee method, function, storage medium and computer program product. BACKGROUND

[0002] In the related art, a wireless network can provide QoS guarantee for an Artificial Intelligence (AI) service, but the technical solution in the related art is susceptible to the influence of network resources, network environment, resource allocation, traffic control, management difficulty and other factors, so that it is difficult to meet the QoS requirement of the high-quality AI service. SUMMARY

[0003] Therefore, the embodiments of the present application aim to provide a QoS guarantee method, function, storage medium and computer program product, which can meet the QoS requirement of the AI service.

[0004] The technical solution of the embodiments of the present application is implemented as follows:

[0005] In a first aspect, the embodiments of the present application provide a QoS guarantee method, applied to a first function, and the method comprises the following steps.

[0006] Receiving first information sent by a second function, the first information being used to describe a requirement related to AI service data transmission and / or a requirement related to AI service computing power;

[0007] Generating a QoS control strategy of the AI service based on the first information and an AI service template corresponding to the subscribed AI service;

[0008] Sending the QoS control strategy to a third function, the QoS control strategy being used for the third function to generate a QoS parameter corresponding to the AI service based on the QoS control strategy.

[0009] In a second aspect, the embodiments of the present application provide a QoS guarantee method, applied to a third function, and the method comprises the following steps.

[0010] Receiving a QoS control strategy sent by a first function, the QoS control strategy being generated by the first function based on first information sent by a second function and an AI service template corresponding to a subscribed AI service;

[0011] Generating a QoS parameter corresponding to the AI service based on the QoS control strategy;

[0012] The QoS parameter is sent to a device related to data transmission and / or a device related to computing power, and the QoS parameter is used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee on the AI service based on the QoS parameter.

[0013] In a third aspect, an embodiment of the present application provides a first function, which includes:

[0014] A first receiving unit is configured to receive first information sent by a second function, and the first information is used to describe a requirement related to AI service data transmission and / or a requirement related to AI service computing power.

[0015] A first processing unit is configured to generate a QoS control policy of the AI service based on the first information and an AI service template corresponding to the AI service in a subscription;

[0016] A first sending unit is configured to send the QoS control policy to a third function, and the QoS control policy is used for the third function to generate a QoS parameter corresponding to the AI service based on the QoS control policy.

[0017] In a third aspect, an embodiment of the present application provides a third function, which includes:

[0018] A second receiving unit is configured to receive a QoS control policy sent by a first function, and the QoS control policy is generated by the first function based on first information sent by a second function and an AI service template corresponding to an AI service in a subscription;

[0019] A second processing unit is configured to generate a QoS parameter corresponding to the AI service based on the QoS control policy.

[0020] A second sending unit is configured to send the QoS parameter to a device related to data transmission and / or a device related to computing power, and the QoS parameter is used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee on the AI service based on the QoS parameter.

[0021] In a fifth aspect, an embodiment of the present application provides a first function, which includes a first processor and a first memory. The first processor implements the QoS guarantee method of the first function side when executing a running program stored in the first memory.

[0022] In a sixth aspect, an embodiment of the present application provides a third function, which includes a second processor and a second memory. The second processor implements the QoS guarantee method of the third function side when executing a running program stored in the second memory.

[0023] In a seventh aspect, an embodiment of the present application provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the first function-side QoS guarantee method or the third function-side QoS guarantee method.

[0024] In an eighth aspect, an embodiment of the present application provides a computer program product comprising a computer program, which, when executed by a processor, implements the first function-side QoS guarantee method or the third function-side QoS guarantee method.

[0025] The embodiments of the present application provide a QoS guarantee method, function, storage medium and computer program product. The method comprises: a first function receiving first information sent by a second function, the first information being used to describe requirements related to AI service data transmission and / or requirements related to AI service computing power; generating a QoS control policy of the AI service based on the first information and an AI service template corresponding to a subscribed AI service; and sending the QoS control policy to a third function, the QoS control policy being used for the third function to generate a QoS parameter corresponding to the AI service based on the QoS control policy. The third function receives the QoS control policy sent by the first function, the QoS control policy being generated by the first function based on the first information sent by the second function and the AI service template corresponding to the subscribed AI service. The third function generates a QoS parameter corresponding to the AI service based on the QoS control policy, and sends the QoS parameter to a device related to data transmission and / or a device related to computing power, the QoS parameter being used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee on the AI service based on the QoS parameter. With the implementation scheme, after the second function transmits the first information to the first function, the first information corresponds to the requirements related to AI service data transmission and the requirements related to AI service computing power, so that the QoS parameter finally generated based on the first information and the AI service template also corresponds to the specific requirements of the AI service. Therefore, when the device related to data transmission and / or the device related to computing power executes the corresponding AI service by using the corresponding QoS parameter, the specific requirements of different AI services on QoS can be met. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 A QoS guarantee method flowchart provided by an embodiment of the present application Figure 1 ;

[0027] Figure 2 An execution process diagram of vehicle networking-intelligent collision prediction provided by an embodiment of the present application

[0028] Figure 3A QoS guarantee method flow provided for an embodiment of the present application Figure 2 ;

[0029] Figure 4 A QoS guarantee method flow provided for an embodiment of the present application

[0030] Figure 5 A QoS guarantee method flow provided for an embodiment of the present application

[0031] Figure 6 A first function component structure provided for an embodiment of the present application Figure 1 ;

[0032] Figure 7 A first function component structure provided for an embodiment of the present application Figure 2 ;

[0033] Figure 8 A third function component structure provided for an embodiment of the present application Figure 1 ;

[0034] Figure 9 A third function component structure provided for an embodiment of the present application Figure 2 . DETAILED DESCRIPTION

[0035] In order to enable a person skilled in the art to more fully understand the features and technical contents of the embodiments of the present application, the technical solutions of the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments of the present application. The accompanying drawings are only used for reference and are not intended to limit the embodiments of the present application.

[0036] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meanings as those commonly understood by a person skilled in the art to which the present application belongs. The terms used in the embodiments of the present application are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0037] In the following description, “some embodiments” are described, which describe a subset of all possible embodiments, but it can be understood that “some embodiments” can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. It should be pointed out that the terms “first / second / third” involved in the embodiments of the present application are only used to distinguish similar objects, and do not represent a specific order of the objects. It can be understood that “first / second / third” can be interchanged with a specific order or sequence as allowed, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0038] In related technologies, the method for wireless network to provide QoS guarantee for AI services usually adopts the following key technical solutions:

[0039] (1) Resource management based on Service Level Agreement (SLA): By establishing SLA, the performance indicators and guarantee requirements of AI services are determined, including delay, bandwidth, availability, etc. Based on SLA, network resources can be dynamically allocated and managed according to different AI service requirements to ensure meeting the quality of service requirements.

[0040] (2) Network slicing technology: Using network slicing technology, the wireless network is divided into multiple virtual slices, each of which can provide dedicated resources for specific AI services. By providing independent network slices for different AI services, resource isolation can be achieved to avoid resource contention and ensure the QoS of AI services.

[0041] (3) Reservation resource strategy: A certain amount of resources are reserved for critical AI services to ensure their priority and performance. The reservation resource strategy can prioritize the scheduling of data transmission and computing tasks for critical AI services in network congestion to ensure the stability and timeliness of the critical AI services.

[0042] (4) Dynamic resource allocation and adjustment: According to the real-time changes of AI service requirements, network resources are dynamically allocated and adjusted. By monitoring network conditions and AI service load in real time, resource scheduling and optimization are performed to meet the quality of service requirements of AI services.

[0043] (5) QoS-aware traffic control: Through traffic control mechanisms, the traffic of each AI service in the network is limited to avoid excessive resource consumption or congestion. Different traffic scheduling algorithms are used for different AI services to optimize network resource utilization and ensure that QoS meets the requirements.

[0044] In summary, through resource management based on SLA, network slicing technology, reservation resource strategy, dynamic resource allocation and adjustment, and QoS-aware traffic control, wireless network can provide QoS guarantee for AI services. These technical methods can dynamically allocate and manage network resources according to service requirements to ensure the performance and reliability of each AI service.

[0045] However, although the method in related technologies can provide QoS guarantee for AI services in wireless network, there are still the following deficiencies, mainly including:

[0046] (1) Resource limitation: The computing resources and bandwidth in wireless network are limited, making it difficult to meet the high-quality AI service requirements. When multiple AI services request resources simultaneously, resource bottlenecks and performance degradation may occur.

[0047] (2) Network environment is difficult to predict: The network environment of wireless networks is complex and changeable, and network delay and bandwidth capacity can be affected by other factors such as network congestion, signal interference, etc. It makes it more difficult to accurately predict and guarantee the QoS of AI services.

[0048] (3) SLA management complexity: Establishing and managing SLA requires negotiation and agreement on various indicators, including delay, bandwidth, availability, etc. It involves coordination and cooperation among multiple parties, and may have problems such as ambiguous interpretation, disputes, etc.

[0049] (4) Difficulty in dynamic resource allocation: Even if dynamic resource allocation and adjustment strategies are adopted, real-time adjustment of network resources according to AI service requirements, implementation will face certain challenges. The accuracy and efficiency of resource allocation are often limited by the performance of real-time monitoring and prediction.

[0050] (5) Complexity of traffic control: In high load situations, achieving good traffic control and scheduling is a complex task. Reasonably allocate the traffic of each AI service to ensure that each service can meet the QoS requirements, which needs to consider various network factors and use intelligent scheduling algorithms.

[0051] To solve the above problems, the embodiments of the present application propose a QoS guarantee method. The wireless network can provide the required computing power, connection resources, etc. for AI services according to the QoS guarantee level of AI services, which can solve the problem that the related art does not have QoS guarantee for AI service related business transmission and QoS guarantee for computing power resources, so as to meet the demand of network edge AI service (English can be expressed as Service) for low service response delay and high service accuracy.

[0052] The technical solution in the embodiments of the present application is implemented as follows. The embodiments of the present application provide a QoS guarantee method, as shown in Figure 1 The method can include the following steps:

[0053] S101, receiving first information sent by a second function.

[0054] The first information is used to describe the demand related to AI service data transmission and / or the demand related to AI service computing power.

[0055] In the embodiments of the present application, the first function and the second function can be referred to as a function body (the first function is referred to as a first function body, and the second function is referred to as a second function body), a communication node (the first function is referred to as a first communication node, and the second function is referred to as a second communication node), or a network element (the first function is referred to as a first network element, and the second function is referred to as a second network element), and the like. In the embodiments of the present application, the names of the first function and the second function are not specifically limited, and different ways can be selected according to the roles of the first function and the second function, as long as the corresponding functions can be implemented.

[0056] In the embodiments of the present application, the first function can include a policy control function (PCF).

[0057] In the embodiments of the present application, the second function can include an application function (AF).

[0058] In the embodiments of the present application, the first function is used to generate a QoS control policy of an AI service, and the second function is used to send requirements related to AI service data transmission and / or requirements related to AI service computing power to the second function.

[0059] In the embodiments of the present application, the first information can be service information (which can be expressed in English as service information).

[0060] In the embodiments of the present application, after a session (which can be expressed in English as session) between a user equipment (UE) and an AF is established / changed, the AF session signaling (which can be expressed in English as session signaling) can carry a session description language (namely, SDI). The AF converts the SDI received by the AF session signaling into service information and delivers the service information to the PCF.

[0061] In the embodiments of the present application, in order to meet the requirements related to AI service data transmission, the data service required by the AI service is taken as a new service, and a media type is added to the media component (MediaComponent) in the session description protocol (Session Description Protocol, SDP). Further, the SDI mapping function, service information, authorized QoS parameters, QoS parameters in the QoS profile and mapping relationship need to be modified accordingly.

[0062] In the embodiments of the present application, in order to meet the requirements related to AI service computing power, the AI task required by the AI service is taken as a new computing component (Computing Component) and added to the SDP. Further, the SDI mapping function, service information, authorized QoS parameters, QoS parameters in the QoS profile and mapping relationship need to be modified accordingly.

[0063] In the embodiments of the present application, in order to meet the requirements related to AI service data transmission and / or the requirements related to AI service computing power, the AF needs to have the function of converting the requirements into service information. Therefore, the function of converting the requirements related to AI service into service information needs to be added to the SDP parameters (SDP parameters) in the AF. Referring to the related content of the SDP parameters in the related art, the values of the SDP parameters sent by an application server are listed in Table 1 as follows.

[0064] In Table 1, V represents the protocol version, o represents the session creator, s represents the session name, i represents the session information, t represents the time information, m represents the media description, c represents the connection information, and a represents the attribute.

[0065] Table 1

[0066]

[0067] In the embodiment of the present application, the requirement related to AI service data transmission in the SDP parameters is added in the AF, which is converted into the function of service information, specifically, the related technology can be referred to, as shown in the following Table 2, which describes the rules for derivation of service information within MediaComponent Description from SDP media component (English can be expressed as Rules for derivation of service information within MediaComponent Description from SDP media component). When the session is started or modified, the AF should derive the media component description AVP of the Rx interface or the “media component” attribute of the N5 interface from the SDP parameters. The SDP parameters are described in IETF RFC (English can be expressed as The SDP parameters are described in IETF RFC).

[0068] Table 2

[0069]

[0070]

[0071] In the embodiment of the present application, the definition of type media component (English can be expressed as Definition of type MediaComponent) can be referred to the related technology, as shown in the following Table 3.

[0072] Table 3

[0073]

[0074]

[0075] Wherein, medType represents a media type parameter, maxSuppBwDl represents a downlink support maximum bandwidth parameter, maxSuppBwUl represents an uplink support maximum bandwidth parameter, and tsnQos represents a transmission QoS parameter.

[0076] In the embodiments of the present application, the enumeration media type (English can be expressed as Enumeration MediaType) defines Enumeration:MediaType is: enumeration "MediaType" represents the media type of a media component (English can be expressed as The enumeration"MediaType"represents the media type of a media component), which can refer to related technologies, such as Table 4 below.

[0077] Table 4

[0078]

[0079]

[0080] In summary, the following content is mainly increased relative to the related art:

[0081] 1) In the Enumeration value of MediaType, AIServiceData and AIServiceModel need to be added, that is, the content added in Table 4 above, that is, the specific content added for the related art is shown in Table 5 below.

[0082] Table 5

[0083]

[0084] 2) In Table 3 of Definition of type MediaComponent, that is, for the related art, AIServiceData / AIServiceModel is added in Applicability, and the QoS parameters corresponding to AIServiceData or AIServiceModel are added in Attribute name. Some of the added content for Table 3 is exemplarily listed in Table 6, which is shown in Table 6 below.

[0085] Table 6

[0086]

[0087]

[0088] It should be noted that the added content in Table 6 is only used to illustrate the position to be modified and the newly added part of the content. Specifically, it can be added according to the actual situation, and the embodiments of the present application do not make specific limitations.

[0089] 3) In Rules for derivation of service information within MediaComponent Description from SDP media component, i.e. for related art, the rules of converting QoS parameters into Service information are defined in Table 2 above.

[0090] Exemplarily, as shown in Table 7 below, the AI max Bw Ul parameter is taken as Service information per Media, and the value of AI Max Requested Bandwidth-UL is converted from the corresponding Derivation from SDP parameter, i.e. the process of taking the corresponding Derivation from SDP parameter can be used to convert Service information.

[0091] It should be noted that, compared with related art, the parameter Max Requested Bandwidth-UL can be a replacement of the original marBwUl parameter.

[0092] Table 7

[0093]

[0094]

[0095]

[0096] In the embodiments of the present application, it is also necessary to add a function of converting the demand of AI service computing task computing power in the SDP parameters into service information in the AF. Specifically, the following content can be added by analogy with the Rules for derivation of service information within Media Component Description from SDP media component involved in Table 2:

[0097] 1) A type Computing Component is added in the SDP.

[0098] 2) Fill in the various types of computing tasks in the Enumeration value of Computing Type, including the training tasks, inference tasks, model selection tasks, model updating tasks, etc. of the AI service computing tasks. As shown in Table 8 below, the added enumeration computing type (English can be expressed as Enumeration ComputingType).

[0099] Table 8

[0100]

[0101] 3) In the Computing Component Type Definition (English can be expressed as Definition of type ComputingComponent) table (see Table 3 above), for different AI service computing tasks as Applicability, make QoS parameters as Attribute name, which can express the QoS parameters that need to be guaranteed by the AI service task (English can be expressed as AI Service task). The table corresponding to the added Definition of type Computing Component is shown in Table 9 below.

[0102] Table 9

[0103]

[0104] Based on the above embodiments, by adding the function of converting the AI service related requirements and the computing power related requirements of the AI service computing task in the SDP parameters into service information in the AF, it is ensured that after the AF and the UE establish a session connection, the AI service related requirements and the computing power related requirements of the AI service computing task can be converted into service information and sent to the PCF for further processing by the PCF.

[0105] S102, generating a QoS control policy of the AI service based on the first information and an AI service template corresponding to the AI service of the subscription.

[0106] In the embodiments of the present application, the AI service template at least includes one or more of the following parameters: data transmission related QoS parameters corresponding to the QoS guarantee level of the AI service and / or computing power related QoS parameters corresponding to the QoS guarantee level.

[0107] In the embodiments of the present application, the QoS parameters related to the computing power corresponding to the QoS guarantee level include at least one or more of the following parameters: computing resource type (expressed in English as Computing Resource type), computing power guaranteed floating point operation per second (Computing_Guaranteed Flops, GFPS), computing power maximum floating point operation per second (Computing_Maximum Flops, MFPS), and computing power priority level (expressed in English as Computing Priority Level).

[0108] In the embodiments of the present application, an AI service template containing AI service and QoS guarantee level information of signing is designed, and the AI service template provides QoS parameters related to computing power and data transmission corresponding to the QoS level. The AI service ID (AI service sub-scene ID or AI service scene ID) and the QoS guarantee level can be managed and numbered according to the scene, which can expand more scenes of AI services that the network can provide, and can subdivide QoS levels for the same AI service to meet more types of user needs.

[0109] In the embodiments of the present application, the AI service user subscribes to the AI service and the QoS guarantee level related to the AI service experience provided by the network to the AI service platform, and the AI service platform queries the corresponding AI service template according to the AI service and the QoS guarantee level subscribed by the user.

[0110] It should be noted that the designation of the AI service template is not limited to standardized formulation, network operator pre-configuration, etc.

[0111] In the embodiments of the present application, the AI service template can be as listed in Tables 10-12.

[0112] Table 10 describes that the user can subscribe to the corresponding AI service use case according to the service scene and the service quality requirement.

[0113] Table 11 describes the QoS parameters related to data transmission corresponding to the AI service QoS guarantee level.

[0114] Table 12 describes the QoS parameters related to computing power resources corresponding to the AI service QoS guarantee level.

[0115] Table 10

[0116] AI service use case number AI service scenario ID AI service sub-scenario ID QoS guarantee level

[0117] Table 11

[0118]

[0119] Table 12

[0120]

[0121] Exemplarily, taking the vehicle networking-smart collision prediction as an example, Figure 2 The corresponding execution process diagram of the vehicle networking-smart collision prediction is shown, and the AI service template corresponding to the vehicle networking-smart collision prediction can be as shown in the following table 13.

[0122] Table 13

[0123]

[0124] For flammable and explosive vehicles, the QoS guarantee level should be 1.1.1, for buses, trucks and other vehicles with blind area, the QoS guarantee level should be 1.1.2, and for ordinary vehicles, the anti-collision information is 1.1.3.

[0125] As shown in the following table 14, the network data transmission QoS requirements of the vehicle networking-smart collision prediction are described in table 14.

[0126] Table 14

[0127]

[0128]

[0129] As shown in the following table 15, the network computing power QoS requirements of the vehicle networking-smart collision prediction are described in table 15.

[0130] Table 15

[0131]

[0132] In the embodiments of the present application, part of the content AI service template listed in the above table 10 to table 12, specifically, the AI service template includes but is not limited to the following content:

[0133] 1) AI service case number: the AI service case number can uniquely represent the correspondence between the AI service and the QoS guarantee level.

[0134] 2) AI service scenario ID: the network pre-numbers according to the AI service application scenarios that can be provided.

[0135] 3) AI service sub-scenario ID: the network can manage and number according to the scene layering, which is conducive to expanding more scenarios of AI services that the network can provide.

[0136] 4) QoS class of protection level: network is classified according to the user experience that can be provided, and the network further designs corresponding QoS parameters according to the QoS protection level.

[0137] Among them, the AI service scenario ID and the QoS protection level can be managed and numbered according to the scene, which has the following advantages: expanding more scenes of AI services that the network can provide; for the same AI service, subdividing the QoS protection level to meet more types of user needs.

[0138] 5) QoS parameters related to AI service data transmission include but are not limited to:

[0139] A, 5G QoS identifier (i.e. 5QI).

[0140] B, ARP.

[0141] C, notification control (English can be expressed as Notification control).

[0142] D, Flow Bit Rates (English can be expressed as Flow Bit Rates) including:

[0143] 1) Guaranteed Flow Bit Rate (GFBR) - DL and UL;

[0144] 2) Maximum Flow Bit Rate (MFBR) - DL and UL.

[0145] E, Aggregate Bit Rates (English can be expressed as Aggregate Bit Rates) including:

[0146] 1) per UE Aggregate Maximum Bit Rate (UE-AMBR);

[0147] 2) per UE per Slice-Maximum Bit Rate (UE-Slice-MBR).

[0148] Among them, the 5G QoS characteristics associated with 5QI include:

[0149] F, Resource type (Non-GBR, GBR, Delay-critical GBR);

[0150] G, Priority Level;

[0151] H. Packet Delay Budget (including Core Network Packet Delay Budget)

[0152] I. Packet Error Rate (Packet Error Rate);

[0153] J. Averaging window (for GBR and Delay-critical GBR resource type only)

[0154] K. Maximum Data Burst Volume (for Delay-critical GBR resource type only)

[0155] Based on the aforementioned QoS parameters related to AI service data transmission, corresponding service assurance QoS parameters related to AI service computing power are designed, including but not limited to the following:

[0156] a. Computing_5QI: The Computing-5QI of computing power is a scalar, and each value corresponds to a defined 5G computing power QoS characteristic parameter (such as d to f below);

[0157] b. Computing_ARP (ARP stands for Allocation and Retention Priority): includes priority, preemption capability, and preemption vulnerability.

[0158] c. Computing_Flops: The number of floating-point operations per second performed by the computing power unit, including:

[0159] 1) Computing power guarantees the number of floating-point operations per second (Computing_Guaranteed Flops, GFPS);

[0160] 2) Computing_Maximum Flops (MFPS)

[0161] The 5G Computing QoS characteristics associated with Computing-5QI include:

[0162] d. Computing_Resource_type: (Non-Guaranteed, Guaranteed, Delay-critical Guaranteed);

[0163] e. Computing_Priority Level: to express the priority of an AI computing task, which can be used to distinguish different AI services of the same user in resource scheduling, or to distinguish different users;

[0164] f. Computing_Averaging window (for Guaranteed, Delay-critical Guaranteed resource type only);

[0165] g. Computing Packet Delay Budget.

[0166] In the embodiments of the present application, since the AI service template contains the QoS guarantee level corresponding to the AI service, the network preconfigures the related QoS parameters (including data transmission related QoS parameters and computing power related QoS parameters) for the AI service according to the QOS level subscribed by the user, establishes an AI service data service detection mechanism, and through the AI service shunting function of the UPF, when different AI services are detected, the network device provides the required connection resources and / or computing power resources for the AI service. When establishing a data bearer (i.e. from the core network to the UE), the data transmission related QoS parameters corresponding to the AI service are used, and when deploying a computing task, the computing power related QoS parameters corresponding to the AI service are used. Specifically, when the UPF data shunting function detects related data services, the data of different services can be distinguished using IP five-tuple or service detection algorithm, and the corresponding data related connection QoS parameters are used to guarantee the service AI; the detection mechanism of the AI task is established, and when the AI task shunting function detects related AI tasks, different AI tasks can be distinguished through AI task related detection algorithm, and the corresponding computing power related QoS parameters are used to guarantee the AI service.

[0167] In the embodiments of the present application, after the network preconfigures the relevant QoS parameters for the AI service, when the PCF receives the service information sent by the AF, the PCF converts the received service information into Authorized QoS parameters / per Service data flow, and the PCF merges the Authorized QoS parameters / per flow (i.e., after merging per Service data flow) in each direction and delivers them to the third function.

[0168] It should be noted that the third function can be referred to as a third function body, a third communication node or a third network element, etc. In the embodiments of the present application, the name of the third function is not limited specifically, and different ways can be selected according to the role of the third function, as long as the corresponding function can be realized.

[0169] In the embodiments of the present application, the third function includes a session management function (SMF).

[0170] In the embodiments of the present application, when the PCF converts the service information into Authorized QoS parameters, the corresponding function is required, and the function of converting the QoS parameters related to the AI service in the SDP parameters into Authorized QoS parameters can be added in the PCF.

[0171] In the embodiments of the present application, the function of the PCF is usually to extract the Authorized 5G QoS Identifier (i.e., 5QI), authorized allocation and retention priority (ARP) and authorized maximum / guaranteed data rate UL / DL (English can be expressed as Authorized Maximum / Guaranteed Data Rate UL / DL).

[0172] Therefore, the function of converting the QoS parameters related to the AI service data transmission in the service information into Authorized QoS parameters can be added in the PCF.

[0173] In the embodiments of the present application, according to the related art, according to the derivation rules (English can be expressed as Rules for derivation of the Maximum Authorized Data Rates, Authorized Guaranteed Data Rates, Maximum Authorized QoS Class and other authorized QoS parameters per service data flow or bidirectional combination of service data flows in the PCF) of the maximum authorized data rate, authorized guaranteed data rate, maximum authorized QoS class and other authorized QoS parameters of each service data flow or bidirectional combination of service data flows in the PCF, for example, the information of the QoS parameters related to the AI service data transmission in the following Table 16 can be added on the basis of the related art, as shown in the following Table 16:

[0174] Table 16

[0175]

[0176]

[0177]

[0178] It should be noted that compared with the Rules for derivation of the Maximum Authorized Data Rates, Authorized Guaranteed Data Rates, Maximum Authorized QoS Class and other authorized QoS parameters per service data flow or bidirectional combination of service data flows in the PCF in the related art, the Max_DR_DL / UL in the related art table can be replaced by AI_Max_DR_DL / UL in the above Table 16, and AImaxBwDl / UL is referred to in the formula of the corresponding column "Derivation from service information", so as to further obtain the Authorized QoS parameters through conversion.

[0179] In the embodiments of the present application, with reference to the related art, according to the calculation rules of the maximum authorized / guaranteed data rate, 5QI and ARP in the PCF (which can be expressed in English as Rules for calculating the Maximum Authorized / Guaranteed Data Rates, 5QI and ARP in the PCF), for example, the information of the QoS parameters related to AI service data transmission in Table 17 can be added on the basis of the related art.

[0180] Table 17

[0181]

[0182] Further, according to Table 17, the authorized QoS parameters of all service data flows defined in the PCC rule, or the authorized QoS parameters of all service data flows in the PDU session, or all service data flows and corresponding AF sessions (which can be expressed in English as all service data flows with corresponding AF session) can be further obtained.

[0183] In the embodiments of the present application, the function of converting the QoS parameters related to the AI service computing task computing power in the service information into the authorized QoS parameters can also be added in the PCF.

[0184] In the embodiments of the present application, with reference to the related art, according to the derivation rules (in English, it can be expressed as Rules for derivation of the Maximum Authorized Data Rates, Authorized Guaranteed Data Rates, Maximum Authorized QoS Class and other authorized QoS parameters per service data flow or bidirectional combination of service data flows in the PCF) of the maximum authorized data rate, authorized guaranteed data rate, maximum authorized QoS class and other authorized QoS parameters of each service data flow or bidirectional combination of service data flows in the PCF, the information of the QoS parameters related to the AI service computing task power can be added on the basis of the related art. Specifically, reference can be made to the related content of AI service data transmission listed in the foregoing embodiments.

[0185] In the embodiments of the present application, with reference to the related art, according to the calculation rules (in English, it can be expressed as Rules for calculating the Maximum Authorized / Guaranteed Data Rates, 5QI and ARP in the PCF) of the maximum authorized / guaranteed data rate, 5QI and ARP in the PCF, the information of the QoS parameters related to the AI service computing task power can be added on the basis of the related art. Specifically, reference can be made to the related content of AI service data transmission listed in the foregoing embodiments.

[0186] Further, all authorized QoS parameters of AI service tasks defined in the PCC rule, or all authorized QoS parameters of service data flows in the PDU session, or all service data flows and corresponding AF sessions (in English, it can be expressed as all service data flows with corresponding AF session) can be further obtained.

[0187] It should be noted that the added parameters related to the AI service power can refer to the adding method of the parameters related to the AI service data transmission, which will not be described here.

[0188] In the embodiments of the present application, the AI service possible AI task type is added in the PCC rule to increase the TaskDetection Rule. The QoS parameters are formulated for each AI Task type according to the QoS level, so that the QoS of the AI computing task required in the AI service can be met as a whole. The AI service possible AI task type includes but is not limited to: training task, inference task, model selection task, model update task, etc.

[0189] S103, send the QoS control policy to the third function.

[0190] The QoS control policy is used for the third function to generate the QoS parameter corresponding to the AI service based on the QoS control policy.

[0191] In the embodiments of the present application, the PCF sends the Authorized QoS parameters to the SMF, and the SMF further generates the QoS parameter corresponding to the AI service according to the Authorized QoS parameters.

[0192] It can be understood that in the QoS guarantee method provided in the embodiments of the present application, after the second function transmits the first information to the first function, the first information corresponds to the data transmission related requirement of the AI service and the computing power related requirement of the AI service, so that the first function generates the QoS control policy of the AI service according to the first information and the AI service template corresponding to the AI service requirement, and then the third function finally generates the QoS parameter corresponding to the specific requirement of the AI service, so that when the corresponding QoS parameter is used to execute the corresponding AI service in the device related to data transmission and / or the device related to computing power, the specific requirement of different AI services to QoS can be met.

[0193] In an embodiment of the present application, after the PCF sends the QoS control policy to the third function, the PCF monitors the QoS information corresponding to the AI service; when it is monitored that the QoS information corresponding to the AI service does not meet the requirement, the network connection resource and the computing power resource at the current time are obtained; based on the network connection resource and the computing power resource, the QoS control policy of the AI service is updated, and the updated QoS control policy is sent to the third function. The updated QoS control policy is used for the third function to generate the updated QoS parameter corresponding to the AI service based on the updated QoS control policy.

[0194] In the embodiments of the present application, the current time can be understood as the time when the QoS information is monitored.

[0195] In the embodiments of the present application, the PCF network element in the network monitors the QoS information of the AI service. When the QoS information of the AI service does not meet the service requirement (for example, the prediction accuracy is low), the network connection resource and the computing power resource at the current time are obtained, and the QoS parameter is selected in the mapping relationship table of the QoS guarantee level and the QoS parameter (that is, the corresponding relationship between one QoS guarantee level and one QoS parameter) according to the monitored QoS information and the network connection and computing power resources, and the AI service template corresponding to the AI service network resource QoS parameter is regenerated, so that the PCF can regenerate the QoS control strategy based on the regenerated QoS parameter.

[0196] It should be noted that the process of regenerating the QoS control strategy can refer to the implementation process of the foregoing embodiments, which will not be described here.

[0197] Exemplarily, taking the vehicle-to-everything-intelligent collision prediction as an example, the QoS parameter in the AI service template is updated according to the network monitored QoS parameter.

[0198] When the network detects that the QoS of the AI service decreases, one of the reasons is that the terminal computing power resource where the AI task 1 (prediction algorithm) is deployed is sufficient, the base station computing power resource where the AI task 2 (decision algorithm) is deployed is sufficient, but the uplink and downlink channel environment of the terminal and the base station is poor, resulting in that the AI service delay is too high.

[0199] In order to guarantee better QoS of the AI service, the QoS parameter in the AI service template can be slightly adjusted, for example, in this case, the QoS parameter of the computing power can be slightly adjusted to reduce the priority of this AI task, and the QoS parameter of the data transmission can be slightly adjusted to improve the data transmission service transmission priority of this AI task, which is more suitable for real-time network resources.

[0200] If referring to the above Tables 13 to 15, for the AI service use case No. 1.1.1-vehicle-to-everything-intelligent collision prediction-1.1.1, the corresponding adjusted QoS requirements are shown in Tables 18-19 below. Table 18 represents the adjusted QoS requirements of the vehicle-to-everything-intelligent collision prediction for network data transmission, and Table 19 represents the adjusted QoS requirements of the vehicle-to-everything-intelligent collision prediction for network computing power.

[0201] Table 18

[0202]

[0203] Table 19

[0204]

[0205] The embodiments of the present application also provide a QoS guarantee method, such as Figure 3As shown, applied to the third function, the method can include:

[0206] S201, receiving the QoS control policy sent by the first function.

[0207] The QoS control policy is generated by the first function based on the first information sent by the second function and the AI service template corresponding to the subscribed AI service.

[0208] In the embodiments of the present application, the first function and the third function have been explained and described in the foregoing embodiments, and will not be repeated here.

[0209] In the embodiments of the present application, the process of generating the QoS control policy by the PCF can refer to the implementation process of the foregoing embodiments, and will not be repeated here.

[0210] S202, generating the QoS parameter corresponding to the AI service based on the QoS control policy.

[0211] In the embodiments of the present application, the QoS parameter includes a data transmission related QoS parameter corresponding to the QoS guarantee level of the AI service and / or a computing power related QoS parameter corresponding to the QoS guarantee level.

[0212] In the embodiments of the present application, the computing power related QoS parameter corresponding to the QoS guarantee level includes at least one or more of the following parameters: computing power resource type, guaranteed floating point operation per second GFPS, maximum floating point operation per second MFPS, and computing power priority.

[0213] In the embodiments of the present application, the SMF receives the authorized QoS parameters and converts the received authorized QoS parameters into access specific QoS parameters.

[0214] In the embodiments of the present application, a function of converting the QoS parameters related to the AI service in the SDP parameters into access specific QoS parameters is added in the SMF.

[0215] Specifically, a function of converting the QoS parameters related to AI service data transmission in the authorized QoS parameters into access specific QoS parameters is added in the SMF.

[0216] In the embodiments of the present application, referring to the related art, the rules for derivation of the authorized QoS parameters per QoS flow from the authorized QoS parameters in SMF (English can be expressed as Rules for derivation of the Authorized QoS Parameters per QoS flow from the Authorized QoS Parameters in SMF) are used to derive the authorized QoS coefficient of each QoS flow according to the authorized QoS parameters in SMF. For example, the information of the QoS parameters related to the AI service data transmission can be added on the basis of the related art, as shown in Table 20 below:

[0217] Table 20

[0218]

[0219]

[0220] In the embodiments of the present application, the function of converting the QoS parameters related to the AI service computing task example in the authorized QoS parameters into the access specific QoS parameters is also needed to be added in the SMF.

[0221] In the embodiments of the present application, referring to the related art, the rules for derivation of the authorized QoS parameters per QoS flow from the authorized QoS parameters in SMF are used to add the information of the QoS parameters related to the AI service computing task on the basis of the related art. The specific way of adding parameters can refer to Table 20 above, which will not be repeated here.

[0222] S203, sending the QoS parameters to the device related to data transmission and / or the device related to computing power.

[0223] The QoS parameters are used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee for the AI service based on the QoS parameters.

[0224] In the embodiments of the present application, the device related to data transmission and / or the device related to computing power can be a device that specifically performs data transmission of the AI service or a device that performs AI service computing tasks, which can be a terminal, a base station, etc. Specifically, it can be selected according to the actual situation, and the present application does not make specific limitations.

[0225] In the embodiments of the present application, the device related to data transmission and / or the device related to computing power receives the corresponding QoS parameter, and performs QoS guarantee on the AI service according to the specific QoS parameter.

[0226] It can be understood that, in the QoS guarantee method provided in the embodiments of the present application, the first information transmitted from the second function to the first function corresponds to the requirements related to data transmission of the AI service and the requirements related to computing power of the AI service, so that the first information transmitted by the third function according to the first information and the AI service template QoS control strategy received by the third function also corresponds to the specific requirements of the AI service, and finally the generated QoS parameter corresponds to the specific requirements of the AI service. Therefore, when the device related to data transmission and / or the device related to computing power uses the corresponding QoS parameter to execute the corresponding AI service, the specific requirements of the AI service on QoS can be met.

[0227] In an embodiment of the present application, the third function can also receive the updated QoS control strategy sent by the first function; based on the updated QoS control strategy, generate updated QoS parameters corresponding to the AI service; and send the updated QoS parameters to the device related to data transmission and / or the device related to computing power. The updated QoS parameters are used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee on the AI service based on the updated QoS parameters.

[0228] In the embodiments of the present application, the PCF sends the updated QoS control strategy to the SMF after updating the QoS control strategy, and the SMF generates updated QoS parameters according to the obtained updated QoS control strategy.

[0229] It should be noted that the way of generating updated QoS parameters according to the obtained updated QoS control strategy can refer to the process of generating QoS parameters according to the QoS control strategy in the foregoing embodiments, and the difference lies in the QoS control strategy. The specific implementation process will not be repeated here.

[0230] In the embodiments of the present application, the SMF sends the QoS parameters to the execution device related to data transmission or the execution device related to AI computing task after generating the updated QoS parameters, and performs QoS guarantee on the execution device related to data transmission or the execution device related to AI computing task.

[0231] Based on the above embodiments, a whole flowchart of QoS guarantee is also provided in the embodiments of the present application, as shown in Figure 4 The following steps are mainly performed:

[0232] 1. After establishing / altering a session, the AF receives the AF session signalling possibly with SID, the AF converts the SID received in the AF session signalling to service information via a SID mapping function and passes it to the PCF.

[0233] 2. The PCF performs a Policy Engine function, the PCF converts the received service information to Authorized QoS parameters / per Service data flow, the PCF merges the Authorized QoS parameters / per flow for each direction and passes it to the SMF.

[0234] 3. The SMF performs a Flow Service Manager function, the SMF converts the received Authorized QoS parameters to access specific QoS parameters and sends it to the UE.

[0235] 4. The SMF sends AI Service detection function rules to the UPF for classifying AI Services for QoS flow marking and other actions.

[0236] 5. The QoS profile is provisioned by the SMF to the gNB or predefined on the gNB.

[0237] 6. The QoS rules are provisioned by the SMF to the UE.

[0238] Figure 4In the specification, SMF is the full name of Session Management Function, which corresponds to the interpretation of session management function; PCF is the full name of Policy Control Function, which corresponds to the interpretation of policy control function; AF is the full name of Application Function, which corresponds to the interpretation of application function; UPF is the full name of User Plane Function, which corresponds to the interpretation of user plane function; UE represents user equipment, and gNB represents next-generation base station.

[0239] Figure 4 In the specification, gNB performs scheduling resources for PDU session (English can be expressed as Schedule resources for PDU session); UE performs mapping UL packets to QoS flows and apply QoS flow marking (English can be expressed as mapping UL packets to QoS flows and apply QoS flow marking).

[0240] Among them, the QoS profile (English can be expressed as profile) transmitted between SMF and gNB needs to be enhanced on the basis of the QoS profile in the related art:

[0241] The QoS profile can be distributed by SMF to gNB, or can be predefined on gNB. In the AN network element of the network, the 5QI related to the connection resources associated with the AI service corresponding scene and its related QoS parameter values are preconfigured.

[0242] In the related art, the 5QI corresponding to the different scenarios of Internet of Vehicles has been defined in the QoS profile, which can be used to guarantee V2X messages (such as advanced driving: collision avoidance (English can be expressed as Advanced Driving: Collision Avoidance)). As shown in the following table 21.

[0243] Table 21

[0244]

[0245]

[0246] In the embodiments of the present application, in the AN network element of the network, the 5QI related to the computing power resources associated with the AI service corresponding scene and its related QoS parameter values are preconfigured, as shown in the following table 22, table 22 shows the enhanced QoS profile.

[0247] Table 22

[0248]

[0249] In the embodiments of the present application, not only QoS guarantee is needed for AI services, but also determination and deployment of AI task or AI subtask deployment scheme corresponding to the AI services are needed.

[0250] In the embodiments of the present application, a whole flowchart for AI service processing is provided, as shown in Figure 5 The main steps include the following:

[0251] 1. AI service subscription and translation, including the following contents:

[0252] 1) User subscribes to AI services;

[0253] 2) AI service template formulation;

[0254] 3) Query AI service template;

[0255] 4) Transmission of AI service template-AI service number.

[0256] 2. AI service arrangement processing, including the following contents:

[0257] 1) AI task disassembly scheme set;

[0258] 2) AI subtask template formulation;

[0259] 3) Transmission of AI service corresponding solution set.

[0260] 3. Optimal solution selection and AI task deployment scheme generation, including the following contents:

[0261] 1) Select the optimal solution;

[0262] 2) Generate AI task deployment scheme;

[0263] 3) Transmission of AI task deployment scheme.

[0264] 4. AI task execution control, including the following contents:

[0265] 1) AI task deployment;

[0266] 2) AI task life cycle management;

[0267] 3) QoS guarantee.

[0268] 5. QoS information monitoring and dynamic adjustment of the optimal solution.

[0269] In the embodiments of the present application, AI service subscription and translation can be performed on the AI service platform to formulate an AI service template. After the AI service template is formulated, the AI service user sends an AI service subscription request to the AI service platform on the network side. After receiving the AI service subscription request, the AI service platform can translate according to the AI service template to obtain an AI service use case number corresponding to the AI service subscription request. Further, the AI service platform completes AI service subscription and sends a subscription success request to the subscribed user. The subscribed user corresponds to the user who sends the AI service subscription request.

[0270] In the embodiments of the present application, the AI service template can include a business template for the AI service and an AI service template for the AI service QoS. The AI service template for the AI service QoS can refer to Tables 10 to 12 in the foregoing embodiments.

[0271] In the embodiments of the present application, the AI service business template includes but is not limited to the following contents:

[0272] 1) Trigger mode of service: start the AI service related business process when the trigger condition is met.

[0273] A, the trigger mode can be event triggering, including but not limited to: positioning detection, weather detection, accident detection, etc.

[0274] B, the trigger mode can be switch triggering, for example, the UE triggers to start the business process by clicking the AI service interface.

[0275] 2) Service network range: the network range that provides service for the user after the AI service business process is triggered.

[0276] 3) AI task ID / description: the functional module required to complete the AI service.

[0277] 4) Input data required by the service.

[0278] 5) Model selection and model source.

[0279] 6) Model task (model selection, model training, model verification optimization, model inference, etc.).

[0280] 7) Model output conclusion / decision, etc.

[0281] 8) AI service case number: the AI service case number can uniquely represent the AI service that a user can subscribe to, and can correspond to the demand of the AI service on the first network resource, thereby translating the AI service that a user can subscribe to into the demand on the network resource. For example, the AI service case number corresponds to a specific AI service, and through the AI service case number, the demand of the specific AI service on the network resource can be determined. Each AI service case number corresponds to the first network resource demand.

[0282] 9) AI service scenario ID: the network (for example, the AI service platform) pre-numbers according to the AI service application scenarios that can be provided.

[0283] 10) AI service sub-scenario ID: the network (for example, the AI service platform) can manage and number according to the scenario layering, which is beneficial to expand more scenarios of the AI services that the network can provide.

[0284] Exemplarily, taking the vehicle networking-intelligent collision detection in Table 23 as an example, after determining that the AI service template ID is 1, the network resource demand of the AI service corresponding to ID 1 can be determined, as shown in Table 24 below.

[0285] Table 23

[0286]

[0287] Table 24

[0288]

[0289] In the embodiment of the present application, the AI service orchestration processing can be specifically: splitting one or more AI tasks in the AI tasks contained in the AI service, for example, splitting between AI function modules (i.e., data acquisition, model inference, model training, model selection, model updating, etc.), or splitting within AI function modules (i.e., the same AI model can be split and executed by different network elements), obtaining a plurality of combined AI sub-tasks, and determining the set composed of the plurality of combined AI sub-tasks and the plurality of AI tasks as a first set.

[0290] In the embodiment of the present application, based on the AI sub-task template, the processing mode corresponding to the AI sub-task and / or the AI task in the first set is determined, and the processing mode corresponding to the AI sub-task and / or the AI task is determined as a second set.

[0291] In the embodiment of the present application, the first set and the second set are used to obtain a third set, and the third set is determined as a solution set of the AI service.

[0292] In the embodiment of the present application, the SMO can perform service orchestration processing on the AI service to obtain a solution set of the AI service.

[0293] In the embodiment of the present application, the first set is all possible combination schemes obtained after the AI task is split.

[0294] In the embodiment of the present application, the second set is a processing manner corresponding to the AI task or the AI subtask.

[0295] In the embodiment of the present application, when the AI task is split to obtain different AI subtasks, the splitting processing manner includes but is not limited to the following splitting processing manners:

[0296] 1) Splitting between AI function modules, such as data collection, model inference, model training, model selection, model updating, etc.

[0297] 2) Splitting within an AI function module, such as splitting the same AI model to be executed by different network elements.

[0298] In the embodiment of the present application, the SMO splits one or more AI tasks in the plurality of AI tasks by using any splitting manner to obtain a plurality of combined AI subtasks, for example, splitting the AI task 1 to obtain the AI task 1 AI subtask 1 and the AI task 1 AI subtask 2, and not splitting the AI task 2, then the AI task 1 AI subtask 1, the AI task 1 AI subtask 2 and the AI task 2 are referred to as one combination.

[0299] In the embodiment of the present application, the first set contains a plurality of combined AI subtasks and a plurality of AI tasks, as shown in Table 25 below.

[0300] Table 25 shows the first set composed of different splitting of AI tasks, wherein the first set only shows three combination manners, but the splitting manner of the AI task is not limited to the following three manners.

[0301] Table 25

[0302]

[0303] In the embodiment of the present application, AI subtask templates of various AI tasks are designed. For example, in the data collection subtask template of the AI task, the data source, the data consumer, the data type, the data task, the data granularity, the data volume, the data processing method, etc. and specific description are included, which are used for the Near-RT RIC on the network side to extract effective information from the data collection subtask template for deployment according to the data collection task, and the Near-RT RIC to perform communication connection resource allocation and guarantee according to the data type, the data task (use), etc. related information.

[0304] In the embodiment of the present application, a data collection subtask template is provided, as shown in Table 26 below:

[0305] Table 26

[0306]

[0307] In the embodiments of the present application, the AI subtask templates of various AI tasks designed also include an AI model task template, and the AI model task template includes one or more of the following: AI task type, AI model, and AI model management related configuration information.

[0308] In the embodiments of the present application, the AI model task template includes one or more of the following: AI task type, AI model, and AI model management related configuration information, which is used to extract valid information from the AI model task template for AI task deployment and life cycle management, and Near-RT RIC performs communication resource and computing resource allocation and guarantee according to the AI task type, AI model and other related information.

[0309] In the embodiments of the present application, the AI model task template is as shown in the following Table 27:

[0310] Table 27

[0311]

[0312] In the embodiments of the present application, the third set obtained by combining the above Tables 25 to 27 is determined as a solution set for solving AI services.

[0313] Exemplarily, the solution set corresponding to the AI service of vehicle networking-intelligent collision prediction can be:

[0314] The AI service of vehicle networking-intelligent collision prediction is completed by AI task 1-service vehicle trajectory prediction and AI task 2-vehicle collision warning decision algorithm, and each task includes data collection, model inference and other processes.

[0315] The network element SMO in the network can generate an AI service solution set by performing service orchestration processing on the AI service. The solution set can include multiple solutions, two of which can be:

[0316] Solution 1: In solution 1, AI task 1 is divided into two subtasks, and the AI service is completed by AI task 1-subtask 1, AI task 1-subtask 2 and AI task 2.

[0317] Among them, AI task 1-subtask 1 is part of the service vehicle trajectory prediction: for example, in a CNN with 3 layers of hidden layer, the inference tasks of the first 2 layers belong to AI task 1-subtask 1.

[0318] AI task 1-subtask 2 is another part of the service vehicle trajectory prediction: as in the CNN, the hidden layer has 3 layers, and the inference task of the last layer belongs to AI task 1-subtask 2.

[0319] AI task 2 is the vehicle collision warning decision algorithm prediction.

[0320] Solution 1 also contains the corresponding AI model task templates, as shown in the following table 28:

[0321] Table 28

[0322]

[0323]

[0324] It should be noted that if data collection is involved in solution 1, the data collection subtask templates in the aforementioned table 26 can be referred to, and examples are not given here.

[0325] Solution 2: In solution 2, the AI service is completed by AI task 1 and AI task 2.

[0326] Solution 2 also contains the corresponding AI model task templates, as shown in the following table 29:

[0327] Table 29

[0328]

[0329] It should be noted that if data collection is involved in solution 2, the data collection subtask templates in the aforementioned table 26 can be referred to, and examples are not given here.

[0330] In the embodiments of the present application, the optimal solution selection and AI task deployment scheme generation specifically includes:

[0331] The SMO of the network side sends the generated solution set of the AI service to the Near-RT RIC to implement the management and arrangement of network resources. When selecting the optimal solution from the solution set, the demand for network resources will also be different due to different AI service solutions. For example: when AI task 1-subtask 1 and AI task 1-subtask 2 are not in the same network element, the intermediate layer data and model parameters need to be transmitted between network elements. When AI task 1 and AI task 2 are not in the same network element: the output data of the vehicle trajectory prediction needs to be transmitted between network elements. Therefore, different AI service solutions correspond to different network resource demands (such as computing resource, connection resource, etc.), and the network resource demand can be computing resource, connection resource, etc.

[0332] Thus, after the SMO sends the solution set to the Near-RT RIC, the Near-RT RIC can determine the network resource requirements corresponding to the AI service solution according to different AI service solutions, and further determine an optimal solution for solving the AI service from the solution set according to the network resource requirements corresponding to each AI service solution.

[0333] In the embodiments of the present application, when selecting an optimal solution from the solution set, for each solution in the solution set, the second network resource required by each solution is estimated according to the information in the AI model task template contained in each solution; the third network resource at the current time is obtained; from the solution set, a solution with the most similar second network resource and third network resource is determined, and the solution with the most similar second network resource and third network resource is determined as the optimal solution.

[0334] In the embodiments of the present application, it is possible that the third network resource does not satisfy all solutions, for example, in the case of extreme resource shortage, a solution closest to / most suitable for / most matched to / most similar to is selected.

[0335] In the embodiments of the present application, the second network resource and the third network resource include one or more of the following: computing power resource, connection resource.

[0336] In the embodiments of the present application, the AI service platform sends the AI service solution set formulated by the AI service platform to the intelligent network controller (i.e., Near-RT RIC) to implement resource management arrangement, that is, according to the network perceived network resource (i.e., the third network resource obtained at the current time), the optimal algorithm is used to select a solution most suitable for the current real-time resource situation, and a deployment scheme of the AI task is further generated.

[0337] The third network resource or the second network resource perceived by the network can include:

[0338] 1) Connection resource corresponding to the coverage range of the AI service:

[0339] a PRB utilization rate of the cell;

[0340] b Wireless channel condition of the cell;

[0341] c Connection load condition of the cell.

[0342] 2) Computing power resource corresponding to the coverage range of the AI service:

[0343] a Hardware structure: GPU, CPU, etc.

[0344] b Computing power size;

[0345] c data storage space.

[0346] The selection of the solution most suitable for the current real-time resource condition is based on cross-domain collaborative management and arrangement of multi-objective optimization algorithms for data, computing power, connection, AI model, etc. The solution set can obtain the demand of different solutions for network resources in advance through information such as AI task type, AI model, etc., including the demand for limited computing power and connection resources of the network. The complex AI service arrangement can be mapped to the matching and selection of network resource demand and real-time sensing resources.

[0347] In the embodiment of the present application, the deployment scheme of the AI task refers to the network element to which the data collection, AI model task, etc. are allocated for specific deployment.

[0348] Exemplarily, taking the AI service optimal solution selection and deployment scheme generation of the vehicle-to-everything-intelligent collision prediction as an example, the resource management and arrangement layer of the network determines that the service vehicle has the inference data and computing power of the AI task 1 according to the perceived network resources; the service vehicle and the main service cell transmission channel quality can meet the real-time requirement of AI task 1 inference result transmission and task 2 decision algorithm issuing; the main service cell has sufficient computing power, storage, etc. to support AI task 2. Therefore, the solution 3 in Table 21 is selected, and the deployment scheme of the AI task is generated, as shown in the following Table 30.

[0349] Table 30

[0350]

[0351] In the embodiment of the present application, after the resource management and arrangement layer of the network (which can be Near-RT RIC) formulates the deployment scheme of the AI task, the life cycle process of the AI service can be executed, and the life cycle of the AI task in the AI service solution is executed, and the QoS guarantee parameters of the AI service are determined to guarantee the QoS of the AI service. The method of QoS guarantee can refer to the foregoing embodiments, which will not be described here.

[0352] In the embodiment of the present application, while guaranteeing the QoS of the AI service, the QoS parameters of the AI service also need to be monitored. When the QoS parameters do not meet the requirements of the AI service, the determination of the QoS parameters is re-performed or when the QoS information does not meet the requirements of the AI service, the optimal solution corresponding to the third network resource is re-selected from the solution set, and the updated AI task deployment scheme is determined based on the optimal solution corresponding to the third network resource.

[0353] Based on the foregoing embodiment, in another embodiment of the present application, a first function 1 is provided, as shown in Figure 6 The first function 1 includes:

[0354] The first receiving unit 10 is configured to receive first information sent by the second function, and the first information is used to describe requirements related to AI service data transmission and / or requirements related to AI service computing power.

[0355] The first processing unit 11 is configured to generate a QoS control policy of the AI service based on the first information and an AI service template corresponding to the AI service in the subscription.

[0356] The first sending unit 12 is configured to send the QoS control policy to the third function, and the QoS control policy is used for the third function to generate a QoS parameter corresponding to the AI service based on the QoS control policy.

[0357] In an embodiment, the first function 1 can further include a monitoring unit and an acquisition unit.

[0358] The monitoring unit is configured to monitor QoS information corresponding to the AI service.

[0359] The acquisition unit is configured to acquire network connection resources and computing power resources at the current time when it is monitored that the QoS information corresponding to the AI service does not meet the requirements.

[0360] The first sending unit 12 is further configured to update the QoS control policy of the AI service based on the network connection resources and the computing power resources, and send the updated QoS control policy to the third function, and the updated QoS control policy is used for the third function to generate an updated QoS parameter corresponding to the AI service based on the updated QoS control policy.

[0361] In an embodiment, the AI service template includes at least one or more parameters of the following: data transmission related QoS parameters corresponding to a QoS guarantee level of the AI service and / or computing power related QoS parameters corresponding to the QoS guarantee level.

[0362] In an embodiment, the computing power related QoS parameters corresponding to the QoS guarantee level include at least one or more parameters of the following: a computing power resource type, a guaranteed floating point operation per second GFPS, a maximum floating point operation per second MFPS, and a computing power priority.

[0363] The embodiment of the application provides a first function, receives first information sent by a second function, the first information is used for describing a requirement related to AI service data transmission and / or a requirement related to AI service computing power; generates a QoS control strategy of the AI service based on the first information and an AI service template corresponding to a subscribed AI service; and sends the QoS control strategy to a third function, the QoS control strategy is used for the third function to generate a QoS parameter corresponding to the AI service based on the QoS control strategy. It can be seen that the first function provided in the embodiment of the application, after the second function transmits the first information to the first function, the first information corresponds to the data transmission related requirement of the AI service and the requirement related to the AI service computing power, so that the first function generates the QoS control strategy of the AI service according to the first information and the AI service template, and the corresponding AI service requirement, and then the QoS parameter finally generated by the third function also corresponds to the specific requirement of the AI service, so that when the corresponding QoS parameter is used to execute the corresponding AI service on the device related to the data transmission and / or the device related to the computing power, the specific requirement of the AI service to the QoS can be met.

[0364] Figure 7 A component structure diagram of a first function 1 provided in the embodiment of the application is shown in the figure. Figure 7 The first function 1 of the embodiment of the application includes a first processor 13, a first memory 14 and a first communication bus 15.

[0365] In the specific embodiment process, the first receiving unit 10, the first processing unit 11, the first sending unit 12, the monitoring unit and the acquisition unit can be implemented by the first processor 13 located on the first function 1, and the first processor 13 can be at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing image processing device (DSPD), a programmable logic image processing device (PLD), a field programmable gate array (FPGA), a CPU, a controller, a microcontroller and a microprocessor. It can be understood that for different devices, the electronic device used to implement the processor function can also be other, and the embodiment of the application does not make specific limitation.

[0366] In the embodiment of the present application, the first communication bus 15 is used to implement connection and communication between the first processor 13 and the first memory 14; when the first processor 13 executes the running program stored in the first memory 14, the following QoS guarantee method is implemented:

[0367] Receive the first information sent by the second function, where the first information is used to describe the requirements related to AI service data transmission and / or the requirements related to AI service computing power; generate a QoS control policy for the AI ​​service based on the first information and the AI ​​service template corresponding to the contracted AI service; send the QoS control policy to the third function, where the QoS control policy is used for the third function to generate QoS parameters corresponding to the AI ​​service based on the QoS control policy.

[0368] In one embodiment, the above-mentioned first processor 13 is also used to monitor the QoS information corresponding to the AI ​​service; when it is monitored that the QoS information corresponding to the AI ​​service does not meet the requirements, the network connection resources and computing power resources at the current moment are obtained; based on the network connection resources and computing power resources, the QoS control policy of the AI ​​service is updated, and the updated QoS control policy is sent to the third function. The updated QoS control policy is used for the third function to generate updated QoS parameters corresponding to the AI ​​service based on the updated QoS control policy.

[0369] In one embodiment, the AI ​​service template includes at least one or more of the following parameters: QoS parameters related to data transmission corresponding to the QoS guarantee level of the AI ​​service and / or QoS parameters related to computing power corresponding to the QoS guarantee level.

[0370] In one embodiment, the QoS parameters related to the computing power corresponding to the QoS guarantee level include at least one or more of the following parameters: computing power resource type, the number of floating-point operations GFPS guaranteed by the computing power per second, the maximum number of floating-point operations MFPS executed by the computing power per second, and computing power priority.

[0371] Based on the above embodiment, a third function 2 is provided in another embodiment of the present application, such as Figure 8 As shown, the third function 2 includes:

[0372] The second receiving unit 20 is used to receive the QoS control policy sent by the first function, where the QoS control policy is generated by the first function based on the first information sent by the second function and the AI ​​service template corresponding to the subscribed AI service.

[0373] The second processing unit 21 is configured to generate QoS parameters corresponding to the AI ​​service based on the QoS control policy.

[0374] The second sending unit 22 is configured to send the QoS parameter to the device related to data transmission and / or the device related to computing power, and the QoS parameter is used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee on the AI service based on the QoS parameter.

[0375] In an embodiment, the second receiving unit 20 is further configured to receive the updated QoS control strategy sent by the first function.

[0376] The second processing unit 21 is further configured to generate an updated QoS parameter corresponding to the AI service based on the updated QoS control strategy.

[0377] The second sending unit 22 is further configured to send the updated QoS parameter to the device related to data transmission and / or the device related to computing power, and the updated QoS parameter is used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee on the AI service based on the updated QoS parameter.

[0378] In an embodiment, the QoS parameter includes a data transmission related QoS parameter corresponding to a QoS guarantee level of the AI service and / or a computing power related QoS parameter corresponding to the QoS guarantee level.

[0379] In an embodiment, the computing power related QoS parameter corresponding to the QoS guarantee level includes at least one or more of the following parameters: a computing power resource type, a guaranteed floating point operation per second GFPS, a maximum floating point operation per second MFPS, and a computing power priority.

[0380] The third function provided in the embodiments of the present application receives the QoS control policy sent by the first function, the QoS control policy is generated by the first function based on the first information sent by the second function and the AI service template corresponding to the subscribed AI service; based on the QoS control policy, the QoS parameter corresponding to the AI service is generated; and the QoS parameter is sent to the device related to data transmission and / or the device related to computing power, the QoS parameter is used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee for the AI service based on the QoS parameter. As can be seen, the third function provided in the embodiments of the present application corresponds to the data transmission related requirement of the first information transmitted by the second function to the first function and the computing power related requirement of the AI service, so that the QoS control policy sent by the first function and received by the third function based on the first information and the AI service template also corresponds to the specific requirement of the AI service, and the finally generated QoS parameter corresponds to the specific requirement of the AI service. Therefore, when the device related to data transmission and / or the device related to computing power executes the corresponding AI service by using the corresponding QoS parameter, the specific requirement of the AI service to QoS can be met.

[0381] Figure 9 The third function 2 provided in the embodiments of the present application is shown in a schematic diagram of a component structure. In actual application, based on the same disclosure concept of the above embodiments, as shown in Figure 9 the third function 2 of the embodiments of the present application includes a second processor 23, a second memory 24 and a second communication bus 25.

[0382] In the specific embodiment process, the above-mentioned second receiving unit 20, second processing unit 21 and second sending unit 22 can be realized by the second processor 23 located on the third function 2, and the above-mentioned second processor 23 can be at least one of ASIC, DSP, DSPD, PLD, FPGA, CPU, controller, microcontroller, microprocessor. It can be understood that for different devices, the electronic devices used to realize the functions of the above-mentioned processors can also be other devices, and the embodiments of the present application do not make specific limitations.

[0383] In the embodiments of the present application, the above-mentioned second communication bus 25 is used to realize the connection and communication between the second processor 23 and the second memory 24; and the above-mentioned second processor 23 realizes the following QoS guarantee method when executing the running program stored in the second memory 24:

[0384] receive a QoS control policy sent by the first function, the QoS control policy being generated by the first function based on the first information sent by the second function and an AI service template corresponding to the AI service of the subscribed AI service; generate a QoS parameter corresponding to the AI service based on the QoS control policy; and send the QoS parameter to a device related to data transmission and / or a device related to computing power, the QoS parameter being used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee for the AI service based on the QoS parameter.

[0385] In an embodiment, the second processor 23 is further configured to receive an updated QoS control policy sent by the first function; generate an updated QoS parameter corresponding to the AI service based on the updated QoS control policy; and send the updated QoS parameter to the device related to data transmission and / or the device related to computing power, the updated QoS parameter being used for the device related to data transmission and / or the device related to computing power to perform QoS guarantee for the AI service based on the updated QoS parameter.

[0386] In an embodiment, the QoS parameter includes a data transmission related QoS parameter corresponding to a QoS guarantee level of the AI service and / or a computing power related QoS parameter corresponding to the QoS guarantee level.

[0387] In an embodiment, the computing power related QoS parameter corresponding to the QoS guarantee level includes at least one or more of the following parameters: a computing power resource type, a guaranteed floating point operation per second GFPS, a maximum floating point operation per second MFPS, and a computing power priority.

[0388] Based on the above embodiments, an embodiment of the present application provides a storage medium having a computer program stored thereon, the computer readable storage medium stores one or more programs, the one or more programs are executable by one or more processors, and are applied in the first function / third function, and the computer program implements the QoS guarantee method as described above.

[0389] Based on the above embodiments, an embodiment of the present application provides a computer program product including a computer program, the computer program is executable by one or more processors, and is applied in the first function / third function, and the computer program implements the QoS guarantee method as described above.

[0390] It should be noted that in the embodiments of the present application, the terms "comprising", "including", or any other variant thereof are intended to cover a non-exclusive inclusion, so that a process, method, article, or apparatus that comprises a list of elements does not include only those elements, but can also include other elements that are not expressly listed or inherent to such process, method, article, or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0391] From the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be realized by means of software and necessary general hardware platforms, of course, they can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a plurality of instructions for making an image display device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) execute the methods described in various embodiments of the present application.

[0392] The above is only a specific implementation of the embodiments of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A quality of service (QoS) guarantee method, characterized by, The method applied to the first function comprises: receiving first information sent by a second function, the first information being used to describe requirements related to AI service data transmission and / or requirements related to AI service computing power; generating a QoS control strategy of the AI service based on the first information and an AI service template corresponding to a signed AI service; sending the QoS control strategy to a third function, the QoS control strategy being used for the third function to generate a QoS parameter corresponding to the AI service based on the QoS control strategy.

2. The method of claim 1, wherein, After the sending of the QoS control strategy to the third function, the method further comprises: monitoring QoS information corresponding to the AI service; when it is monitored that the QoS information corresponding to the AI service does not meet the requirements, obtaining network connection resources and computing power resources at the current time; updating the QoS control strategy of the AI service based on the network connection resources and the computing power resources, and sending the updated QoS control strategy to the third function, the updated QoS control strategy being used for the third function to generate an updated QoS parameter corresponding to the AI service based on the updated QoS control strategy.

3. The method of claim 1, wherein, The AI service template comprises at least one or more parameters of a data transmission related QoS parameter corresponding to a QoS guarantee level of the AI service and / or a computing power related QoS parameter corresponding to the QoS guarantee level.

4. The method of claim 3, wherein, The computing power related QoS parameter corresponding to the QoS guarantee level comprises at least one or more parameters of a computing power resource type, a guaranteed floating point operation per second GFPS, a maximum floating point operation per second MFPS, and a computing power priority.

5. A QoS guarantee method characterized by, The method applied to the third function comprises: receiving a QoS control strategy sent by a first function, the QoS control strategy being generated by the first function based on first information sent by a second function and an AI service template corresponding to a signed AI service; generating a QoS parameter corresponding to the AI service based on the QoS control strategy; sending the QoS parameter to a device related to data transmission and / or a device related to computing power, the QoS parameter being used for the device related to the data transmission and / or the device related to the computing power to guarantee QoS of the AI service based on the QoS parameter.

6. The method of claim 5, wherein, The method further comprises: receiving an updated QoS control strategy sent by the first function; generating an updated QoS parameter corresponding to the AI service based on the updated QoS control strategy; sending the updated QoS parameter to the device related to the data transmission and / or the device related to the computing power, the updated QoS parameter being used for the device related to the data transmission and / or the device related to the computing power to guarantee QoS of the AI service based on the updated QoS parameter.

7. The method according to claim 5 or 6, characterized in that, The QoS parameters include a data transmission related QoS parameter corresponding to a QoS guarantee level of the AI service and / or a computing power related QoS parameter corresponding to the QoS guarantee level.

8. The method of claim 7, wherein, The computing power related QoS parameter corresponding to the QoS guarantee level at least includes one or more of the following parameters: a computing power resource type, a guaranteed floating point operation per second GFPS, a maximum floating point operation per second MFPS, and a computing power priority.

9. A first function, characterized by The first function includes: The first receiving unit is configured to receive first information sent by the second function, the first information being used to describe requirements related to AI service data transmission and / or requirements related to AI service computing power. The first processing unit is configured to generate a QoS control policy of the AI service based on the first information and an AI service template corresponding to the subscribed AI service. The first sending unit is configured to send the QoS control policy to a third function, the QoS control policy being used for the third function to generate QoS parameters corresponding to the AI service based on the QoS control policy.

10. A third function, characterized by The third function includes: The second receiving unit is configured to receive a QoS control policy sent by the first function, the QoS control policy being generated by the first function based on the first information sent by the second function and an AI service template corresponding to the subscribed AI service. The second processing unit is configured to generate QoS parameters corresponding to the AI service based on the QoS control policy. The second sending unit is configured to send the QoS parameters to a device related to data transmission and / or a device related to computing power, the QoS parameters being used for the device related to the data transmission and / or the device related to the computing power to perform QoS guarantee for the AI service based on the QoS parameters.

11. A first function, characterized by The first function includes a first processor and a first memory; and the first processor implements the method in any one of claims 1 to 4 when executing a running program stored in the first memory.

12. A second function, characterized by The second function includes a second processor and a second memory; and the second processor implements the method in any one of claims 5 to 8 when executing a running program stored in the second memory.

13. A storage medium having stored thereon a computer program, characterized in that The computer program, when executed by a processor, implements the method in any one of claims 1 to 4, or the computer program, when executed by a processor, implements the method in any one of claims 5 to 8.

14. A computer program product comprising a computer program, characterized in that, The computer program, when executed by a processor, implements the method in any one of claims 1 to 4, or the computer program, when executed by a processor, implements the method in any one of claims 5 to 8.

Images