Handover method and apparatus, network device, storage medium and computer program product

Abstract

The application discloses a switching method and device, a network device, a storage medium and a computer program product. The method comprises the following steps: a first request is sent to a second network device; wherein the first request carries first information, and is used for requesting to switch an AI service provided by a first network device to the second network device; and the first information represents the related requirements of the AI service.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a switching method, apparatus, network device, storage medium, and computer program product. Background Technology

[0002] In related technologies, in scenarios where networks provide artificial intelligence (AI) services to mobile user equipment (UE), the lack of design for switching computing resources makes it impossible to guarantee the quality of AI services, the continuity and stability of wireless connections, and the real-time performance and accuracy of AI calculations. Summary of the Invention

[0003] To address the related technical issues, embodiments of this application provide a switching method, apparatus, network device, storage medium, and computer program product.

[0004] The technical solution of this application embodiment is implemented as follows:

[0005] This application provides a switching method applied to a first network device, the method comprising:

[0006] Send a first request to a second network device; wherein the first request carries first information for requesting to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0007] In the above scheme, the first information includes one or more of the following information from the AI ​​service:

[0008] AI models;

[0009] Calculate the container's configuration parameters;

[0010] AI Service Quality (QoAIS).

[0011] In the above scheme, sending the first request to the second network device includes:

[0012] A first request is sent to the second network device if the signal quality of the second network device meets the handover requirements and one or more of the following conditions are met:

[0013] The second network device supports the AI ​​service;

[0014] The first network device receives first data related to the AI ​​service sent by the first terminal;

[0015] The first network device has not yet started executing the computational tasks related to the AI ​​service;

[0016] The first network device has not completed the computational tasks related to the AI ​​service;

[0017] The remaining computational load of the AI ​​service in the first network device is greater than or equal to the first threshold.

[0018] In the above scheme, before sending the first request to the second network device, the method further includes:

[0019] Send a second request to the second network device, the second request being used to request information related to computing resources;

[0020] The system receives second information sent by the second network device; the second information represents information related to the computing resources of the second network device.

[0021] The method in the above scheme further includes:

[0022] Based on the second information, it is determined whether the second network device supports or does not support the AI ​​service.

[0023] The method in the above scheme further includes:

[0024] Receive a first message sent by the second network device, the first message indicating agreement to switch;

[0025] Send a third message to the second network device, the third message representing information for performing the AI ​​service or the calculation result of the AI ​​service.

[0026] In the above scheme, the third information includes one or more of the following:

[0027] The AI ​​model related to the AI ​​service;

[0028] The input data for the AI ​​service;

[0029] Intermediate result data of the computational tasks related to the AI ​​service;

[0030] The final result of the AI ​​service.

[0031] In the above scheme, the third information is carried in the extension header of the user plane protocol.

[0032] In the above scheme, the second data or calculation results related to the AI ​​service are transmitted through the user plane connection, and the requests and confirmation messages related to the AI ​​service are transmitted through the control plane connection.

[0033] This application embodiment also provides a switching method applied to a second network device, the method comprising:

[0034] Receive a first request sent by a first network device; wherein the first request carries first information for requesting to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0035] In the above scheme, the first information includes one or more of the following information from the AI ​​service:

[0036] AI models;

[0037] Calculate the container's configuration parameters;

[0038] QoAIS.

[0039] In the above scheme, before receiving the first request sent by the first network device, the method further includes:

[0040] Receive a second request sent by the first network device, the second request being used to request information related to computing resources;

[0041] Send a second message to the first network device; the second message represents information related to the computing resources of the second network device.

[0042] In the above scheme, after receiving the first request sent by the first network device, the method further includes:

[0043] Based on the first information, decide whether to respond to or reject the first request;

[0044] If the decision is made to respond to the first request, a first message is sent to the first network device, the first message indicating agreement to the handover.

[0045] The method in the above scheme further includes:

[0046] The system receives third information sent by the first network device, the third information representing information used to perform the AI ​​service or the calculation result of the AI ​​service.

[0047] The method in the above scheme further includes:

[0048] If the second network device meets the relevant requirements of the AI ​​service represented by the first information, a switching preparation is performed based on the configuration parameters of the computing container related to the AI ​​service.

[0049] The method in the above scheme further includes:

[0050] If the second network device meets the relevant requirements of the AI ​​service represented by the first information, a switching preparation is performed based on the configuration parameters of the computing container related to the AI ​​service.

[0051] In the above scheme, the third information includes one or more of the following:

[0052] The AI ​​model related to the AI ​​service;

[0053] The input data for the AI ​​service;

[0054] Intermediate result data of the computational tasks related to the AI ​​service;

[0055] The final result of the AI ​​service.

[0056] In the above scheme, the third information is carried in the extension header of the user plane protocol.

[0057] This application also provides a switching device, including:

[0058] The first sending unit is configured to send a first request to the second network device; wherein the first request carries first information, which is used to request the switching of the artificial intelligence (AI) service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0059] This application also provides a switching device, including:

[0060] The first receiving unit is configured to receive a first request sent by a first network device; wherein the first request carries first information, which is used to request switching the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0061] This application embodiment also provides a first network device, including: a first processor and a first communication interface; wherein,

[0062] The first communication interface is used to send a first request to the second network device; wherein the first request carries first information, which is used to request to switch the artificial intelligence (AI) service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0063] This application also provides a second network device, including: a second processor and a second communication interface; wherein,

[0064] The second communication interface is used to receive a first request sent by the first network device; wherein the first request carries first information for requesting to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0065] This application also provides a network device, including a processor and a memory for storing computer programs that can run on the processor.

[0066] When the processor runs the computer program, it executes the steps of any of the methods described above on the first network device side, or executes the steps of any of the methods described above on the second network device side.

[0067] This application embodiment also provides a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of any of the methods described above for the first network device side, or implements the steps of any of the methods described above for the second network device side.

[0068] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above methods.

[0069] In the switching method, apparatus, network device, storage medium, and computer program product provided in this application embodiment, a first network device sends a first request to a second network device; the second network device receives the first request sent by the first network device; wherein, the first request carries first information for requesting to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service. In this scheme, by sending a first request to the second network device, the first network device can achieve the switching of the network device providing the AI ​​service, including the switching of computing resources, thereby ensuring the continuity and stability of the wireless connection while also ensuring the real-time performance and accuracy of AI computing, as well as the quality of the AI ​​service. Attached Figure Description

[0070] Figure 1 This is a schematic diagram of the handover process between base stations in related technologies;

[0071] Figure 2 This is a flowchart illustrating a switching method according to an embodiment of this application;

[0072] Figure 3 This is a schematic diagram of a network architecture for AI services according to an embodiment of this application;

[0073] Figure 4 This is a flowchart illustrating another switching method according to an embodiment of this application;

[0074] Figure 5 This is a flowchart illustrating a switching method according to an application embodiment of this application;

[0075] Figure 6 This is a schematic diagram of a switching device structure according to an embodiment of this application;

[0076] Figure 7 This is a schematic diagram of another switching device structure according to an embodiment of this application;

[0077] Figure 8 This is a schematic diagram of the structure of the first network device according to an embodiment of this application;

[0078] Figure 9 This is a schematic diagram of the structure of the second network device in the embodiment of this application. Detailed Implementation

[0079] In related technologies, during handover between base stations in a mobile UE scenario, the base station performs measurement configuration. The terminal performs measurements according to the requirements of the measurement configuration and reports the measurement results. The base station makes a handover decision based on the received measurement results. When connection resources change, the base station makes a handover decision based on information such as signal strength and quality measured and reported by the UE, the UE's moving speed and direction, load balancing, and Quality of Service (QoS). Figure 1 As shown, the handover process between base stations includes:

[0080] Step 0: The source base station (also known as the source gNB) transmits mobility control information provided by the access and mobility management function (AMF) between the user plane function (UPF).

[0081] Step 1: The source base station configures the measurement for the UE, and the UE performs measurements based on the measurement configuration and reports the measurement report.

[0082] Step 2: The source base station makes a handover decision based on the measurement report and Radio Resource Management (RRM) information.

[0083] Step 3: The source base station sends a handover request to the target base station (also known as the destination base station or target gNB), transmitting a transparent Radio Resource Control (RRC) container with necessary information to prepare for the handover on the target base station side.

[0084] Step 4: Admission control can be performed by the target base station. If fragment information is sent to the target base station, fragment-aware admission control should be performed; if a Packet Data Unit (PDU) session is associated with an unsupported slice, the target base station should reject such PDU sessions.

[0085] Step 5: The target base station prepares for handover using L1 / L2 and sends a handover request confirmation message (also known as HANDOVERREQUEST ACKNOWLEDGE) to the source base station, which includes a transparent container to be sent to the UE as an RRC message to perform the handover.

[0086] Step 6: After receiving the handover request confirmation message, the source base station triggers Uu handover by sending a Radio Access Network Handover Initiation message (also known as RAN Handover Initiation) to the UE. This message contains the information required to access the target cell, such as the target cell ID, the new C-RNTI, and the selected target base station security algorithm identifier.

[0087] Step 7: For a DRB with DAPS configured, the source base station sends a first status message (such as EARLYSTATUS TRANSFER) to the target base station; for a DRB without DAPS configured, the source base station sends a second status message (such as SNSTATUS TRANSFER) to the target base station to convey the uplink PDCP SN receiver status and downlink PDCP SN transmitter status of the DRB that are eligible for PDCP status reservation.

[0088] Step 8: The UE synchronizes with the target cell and completes the RRC handover process by sending a Radio Access Network (RAB) handover completion message (e.g., RAN Handover Completion) to the target base station. In the case of DAPS handover, the target base station also sends a handover success message (e.g., HANDOVER SUCCESS) to the source base station to notify the UE that it has successfully accessed the target cell. In return, the source base station sends a DRB status transmission message for DAPS-configured DRBs to the target base station. For DRBs configured with DAPS, the source base station will not stop transmitting uplink QoS flows to the UPF until a third status message (e.g., SN STATUSTRANSFER) is sent in step 8b.

[0089] Step 9: The target base station sends a path switching request (such as PATH SWITCH REQUEST) to the AMF to trigger the core network (5GC, 5th Generation Core Network) to switch the DL data path to the target base station and establish a control plane interface (such as NG-C interface) instance between the AMF and the target base station.

[0090] Step 10: UPF switches downlink paths. The 5GC switches the DL data path to the target base station. The UPF sends one or more "end marker" packets to the source base station on the old path of each PDU session / tunnel, and can then release any user plane or transport network layer (TNL) resources to the source base station.

[0091] Step 11: The AMF sends a path switching request confirmation message (such as PATH SWITCH REQUESTACKNOWLEDGE) to the target base station.

[0092] Step 12: Upon receiving the path handover request confirmation message, the target base station sends a context release message (such as UE CONTEXT RELEASE) to the source base station to notify the source base station that the handover was successful. Then, the source base station releases the radio and user plane related resources associated with the UE context.

[0093] When the network provides Artificial Intelligence (AI) services to the User Equipment (UE) (such as real-time AI inference services for image recognition in typical scenarios of mobile robots and autonomous driving), the UE, being mobile, faces frequent intra-base station handovers, inter-base station handovers, and cross-TA handovers. Therefore, ensuring the quality of AI services, the continuity and stability of the connection, and the real-time performance and accuracy of computation during handover is a crucial issue. Currently, terminals consider the quality of the wireless connection when reporting measurements, but do not consider the handover process and judgment criteria for AI computing resources. Therefore, the current system lacks a design for computation handover. If the computing resources of each base station are also acquired by the terminal like wireless resources, the terminal's air interface signaling will incur significant overhead, which will also severely impact the real-time performance of AI services.

[0094] Based on this, in various embodiments of this application, a first network device sends a first request to a second network device; the second network device receives the first request sent by the first network device; wherein, the first request carries first information for requesting the switching of the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service. In the above scheme, by sending a first request to the second network device, the first network device can achieve the switching of the network device providing the AI ​​service, including the switching of computing resources, thereby ensuring the continuity and stability of the wireless connection while also ensuring the real-time performance, accuracy, and quality of the AI ​​service.

[0095] The present application will now be described in further detail with reference to the accompanying drawings and embodiments.

[0096] This application provides a handover method applied to a first network device, which can be understood as a source base station, i.e., a base station currently connected to a first terminal and providing connection and AI services to the first terminal, where the first terminal is any terminal in a mobile scenario; Figure 2 As shown, the method includes:

[0097] Step 201: Send the first request to the second network device.

[0098] The first request carries first information, which is used to request the switching of the AI ​​service provided by the first network device to the second network device. The first information represents the relevant requirements of the AI ​​service.

[0099] Here, in the scenario of a mobile first terminal, i.e., when the first terminal moves from the coverage area of ​​one network device to the coverage area of ​​another network device, a first handover is triggered. The first handover includes a connection handover and an AI service handover. The first request may include a connection handover request, or a connection handover request and an AI service handover request. The connection handover request can be used to request a switch from the first network device to the second network device, i.e., to request the establishment of a communication connection between the first terminal and the second network device. The AI ​​service handover request can be used to request a switch of the AI ​​service provided by the first network device to the first terminal from the first network device to the second network device. If the first request includes a connection handover request, the connection handover request can also trigger an AI service handover request; that is, the connection handover request can also be used to request a switch of the AI ​​service provided by the first network device to the second network device. The second network device can be understood as a target base station or destination base station, with the computing resources of different network devices deployed in different physical locations. The AI ​​service handover request can also be understood as a computing resource handover request, an AI computing resource handover request, or an AI computing handover request.

[0100] It should be noted that the first network device may send a first request to the second network device if it decides or determines to switch the connection of the first terminal from the first network device to the second network device; or it may send a first request directly to the second network device, and the second network device may decide or determine whether to switch the connection of the first terminal from the first network device to the second network device.

[0101] It should be noted that, when the first network device completes the computational tasks related to the AI ​​service and obtains the final result of the AI ​​service, the final result can be transmitted via connection packet data transmission or via the AI ​​service data forwarding function. When the final result is transmitted via connection packet data transmission, the first request can be a connection switching request, used to request switching the first network device connected to the first terminal to the second network device, followed by the execution of connection switching steps. When the final result is transmitted via the AI ​​service data forwarding function, the first request can include both a connection switching request and an AI service switching request. Connection packet data includes data that the first network device has not yet transmitted to the first terminal or has not yet completed transmitting at the time of connection switching; connection packet data can also be called link packet data.

[0102] To ensure a smooth AI service switchover and thus guarantee the real-time performance, accuracy, and quality of the AI ​​service, in one embodiment, the first information includes one or more of the following information about the AI ​​service:

[0103] AI models;

[0104] Calculate the container's configuration parameters;

[0105] QoAIS.

[0106] Here, where the first request includes an AI service switching request, or can be used to request switching the AI ​​service provided by the first network device to the first terminal to the second network device, the first information may include one or more of the following: AI model, computing container configuration parameters, and QoAIS. The AI ​​model represents the AI ​​model that needs to be used or invoked when executing the AI ​​service, or represents the AI ​​model requested by the first terminal, for the second network device to continue providing the AI ​​service; the computing container configuration parameters are used by the second network device to configure the computing container of the AI ​​model to continue providing the AI ​​service; QoAIS may at least include AI service latency and / or AI service accuracy to ensure the quality of the AI ​​service. The computing container includes the AI ​​model and the AI ​​model's runtime environment, obtained by packaging the AI ​​model and the runtime environment into a separate container.

[0107] The first information may include QoAIS, and may also include the configuration parameters of the computing container and / or the AI ​​model; if the first information does not include the configuration parameters of the computing container and / or the AI ​​model, the second network device may contain a computing container that executes the AI ​​model and the AI ​​model, or the QoAIS and the configuration parameters of the computing container and / or the AI ​​model may be sent separately. That is, upon receiving a third request sent by the second network device, the configuration parameters of the computing container and / or the AI ​​model may be sent to the second network device, and the third request is used to request the configuration parameters of the computing container and / or the AI ​​model.

[0108] It should be noted that there may be situations where a neighboring network device receives an AI service request but does not have or support the corresponding AI service model and requests the first network device to cooperate. In the event of receiving a third request from a neighboring network device, the first information can also be sent to the neighboring network device.

[0109] To ensure a smooth switchover of network devices and / or to guarantee the continuity and accuracy of AI services, in one embodiment, sending the first request to the second network device includes:

[0110] A first request is sent to the second network device if the signal quality of the second network device meets the handover requirements and one or more of the following conditions are met:

[0111] The second network device supports the AI ​​service;

[0112] The first network device receives first data related to the AI ​​service sent by the first terminal;

[0113] The first network device has not yet started executing the computational tasks related to the AI ​​service;

[0114] The first network device has not completed the computational tasks related to the AI ​​service;

[0115] The remaining computational load of the AI ​​service in the first network device is greater than or equal to the first threshold.

[0116] Here, when switching connections, it is necessary to consider whether the signal quality of the first terminal meets the requirements for switching connections. When switching AI services, it is necessary to consider whether the computing resources of the second network device support the AI ​​service and whether the current state of the AI ​​service is suitable for switching. The first network device can send a second request to the second network device to obtain the computing resource information of the second network device when making a switching decision, or it can send a second request to the second network device in advance to obtain the computing resource information of the second network device before making a switching decision. For example, after receiving the measurement report of the first terminal, a second request can be sent to the second network device. It can also obtain the latest computing resource information of the second network device periodically.

[0117] Sending a first request to the second network device when certain conditions are met can be understood as follows: Sending a first request to the second network device when the first network device determines that connection switching and AI service switching are permissible; sending a first request to the second network device when the signal quality of the second network device meets the switching requirements and the second network device supports AI services; or sending a first request to the second network device when the signal quality of the second network device meets the switching requirements and the second network device supports AI services, and one or more of the following conditions:

[0118] The first network device receives the first data related to the AI ​​service sent by the first terminal;

[0119] The first network device has not yet started executing AI service-related computing tasks;

[0120] The first network device has not completed the computational tasks related to the AI ​​service;

[0121] The remaining computational load for AI services in the first network device is greater than or equal to the first threshold.

[0122] Among these, the signal quality of the second network device meeting the handover requirements indicates that the second network device meets the connection handover requirements; the second network device supporting AI services indicates that the second network device can execute AI service-related computational tasks, or can continue to execute AI service-related computational tasks; the first data related to the AI ​​service includes the input data of the AI ​​model, or the result data obtained by the first terminal after executing the terminal-side computational task of the AI ​​model, that is, the first data related to the AI ​​service can be understood as the input data of the AI ​​model when the first network device executes the AI ​​service; the first network device receiving the first data related to the AI ​​service sent by the first terminal indicates that the first network device has received the input data of the AI ​​model, but has not yet started executing the AI ​​service-related computational task or The following statements indicate that the first network device has not completed the AI ​​service-related computational tasks: The first network device has not started executing AI service-related computational tasks, indicating that it needs to execute AI service-related computational tasks but has not yet started; the first network device has not completed the AI ​​service-related computational tasks, indicating that it has started executing AI service-related computational tasks and has reached an intermediate node of the AI ​​model or an intermediate step of the AI ​​service-related computational tasks, but has not yet completed all the AI ​​computational tasks that the first network device needs to execute; the remaining computational amount of the AI ​​service in the first network device is greater than or equal to a first threshold, indicating that the first network device has completed a small portion of the AI ​​service-related computational tasks; the first threshold can be set according to the actual application, for example, the first threshold can be set to 20%.

[0123] Considering that switching AI services when the remaining computational load of the AI ​​service in the first network device is less than a first threshold (i.e., when the first network device is about to complete the AI ​​service-related computational tasks), would increase the signaling interaction between the first and second network devices (including the transmission of AI models) and increase the computational complexity of the second network device in executing AI service-related computational tasks (including the computational complexity of resource allocation), the following applies: If the second network device does not support AI services, or if the first network device is executing AI services and the remaining computational load of the AI ​​service is less than the first threshold, the first network device can complete the AI ​​service-related computational tasks first and continue normal data forwarding as connection data. This involves switching the connection and transmitting data that has not yet been transmitted to the first terminal or has not been fully transmitted to the second network device. This eliminates the need for AI service switching, related signaling interaction, and reconfiguration of computational bearers, thus saving signaling overhead and reducing computational complexity.

[0124] It should be noted that when the first request is a connection switching request, there is no need to make an AI service switching decision. That is, the first request can be sent to the second network device as long as the signal quality of the second network device meets the switching requirements.

[0125] To ensure that the network device receiving the first request supports the AI ​​service, thereby improving the success rate of AI service switching, in one embodiment, before sending the first request to the second network device, the method further includes:

[0126] Send a second request to the second network device, the second request being used to request information related to computing resources;

[0127] The system receives second information sent by the second network device; the second information represents information related to the computing resources of the second network device.

[0128] Here, the second request can be sent to the second network device before the first network device executes the handover decision; alternatively, the second request can be sent to the second network device while the first network device is executing the handover decision. That is, when determining whether the second network device supports AI services, information about the second network device's computing resources is needed, and a second request is sent to the second network device. Upon receiving the second information, a handover decision is made; or, if the second information is received during the handover decision process, the second information is used to determine whether the second network device supports AI services, thereby making an AI service handover decision.

[0129] The second information may include parameters indicating the idle computing resources of the second network device, such as the percentage of idle computing resources, and may also include one or more parameters measuring the computing power of the network device. Parameters measuring the computing power of the network device may include one or more of the following: the computing power of the central processing unit (CPU), the computing power of the graphics processing unit (GPU), the computing power of AI-specific acceleration hardware, memory capacity and speed, storage capacity and speed, and energy efficiency. AI-specific acceleration hardware includes one or more of the following: a tensor processing unit (TPU), a field-programmable gate array (FPGA), and an application-specific integrated circuit (ASIC).

[0130] It should be noted that, upon receiving the second information, the percentage of idle computing resources of the second network device can be calculated based on the second information, thereby determining whether the second network device can provide AI services.

[0131] It should be noted that if the first network device obtains relevant information about the computing resources of the second network device periodically, it can make a handover decision directly without sending a second request to the second network device to obtain the second information.

[0132] In order to make a handover decision and complete the handover of network devices, in one embodiment, the method further includes:

[0133] Based on the second information, it is determined whether the second network device supports or does not support the AI ​​service.

[0134] Here, upon receiving the second information, based on the second information, it can be determined whether the computing power of the second network device can meet the computing requirements of the AI ​​service, and also whether the idle computing resources of the second network device meet the computing resource requirements of the AI ​​service; if the computing power of the second network device meets the computing requirements of the AI ​​service, and the idle computing resources of the second network device meet the computing resource requirements of the AI ​​service, it is determined that the second network device supports the AI ​​service; if the computing power of the second network device does not meet the computing requirements of the AI ​​service, or the idle computing resources of the second network device do not meet the computing resource requirements of the AI ​​service, it is determined that the second network device does not support the AI ​​service.

[0135] To enable AI service switching and ensure the continuity of AI services, in one embodiment, the method further includes:

[0136] Receive a first message sent by the second network device, the first message indicating agreement to switch;

[0137] Send a third message to the second network device, the third message representing information for performing the AI ​​service or the calculation result of the AI ​​service.

[0138] Here, after sending a first request to the second network device, receiving a first message from the second network device indicates that the connection of the first terminal can be switched from the first network device to the second network device, and / or the AI ​​service provided by the first network device can be switched to the second network device, i.e., triggering an air interface connection switch and / or an AI service switch; a message is synchronized to the first terminal, i.e., a message notifying the first terminal to perform a connection switch and / or an AI service switch, and relevant data is synchronized to the second network device. The first message can be understood as a response to the first request; if the first request is a connection switch request, the first message triggers an air interface connection switch; if the first request includes both a connection switch request and an AI service switch request, the first message triggers both an air interface connection switch and an AI service switch. The first message may also carry a third request, which can be understood as an AI model sharing request or an AI model acquisition request, so that the second network device can obtain the AI ​​model and thus continue to provide AI services and execute the relevant computational tasks of the unfinished AI services; correspondingly, the third information may also include the AI ​​model.

[0139] It should be noted that if there is a computing load between the first network device and the first terminal when the first message is received from the second network device, then a reconfiguration of the computing load is required.

[0140] It should be noted that when the first network device provides multiple AI services to the first terminal, it can send third information corresponding to different AI services to the second network device.

[0141] It should be noted that the first network device can also send a fourth message to the second network device; the fourth message represents the connection packet data that the first network device has cached and needs to be transmitted to the first terminal or has not yet been transmitted; when the first network device has completed the relevant calculation tasks of the AI ​​service and obtained the final result of the AI ​​service, the fourth message can also include the final result of the AI ​​service, that is, the final result of the AI ​​service is transmitted through the connection packet data.

[0142] In order to synchronize AI service-related data and complete the AI ​​service switch, in one embodiment, the third information includes one or more of the following:

[0143] The AI ​​model related to the AI ​​service;

[0144] The input data for the AI ​​service;

[0145] Intermediate result data of the computational tasks related to the AI ​​service;

[0146] The final result of the AI ​​service.

[0147] Here, when the first network device receives the first data related to the AI ​​service sent by the first terminal (including the input data of the AI ​​model, or the result data obtained after the first terminal executes the terminal-side computation task of the AI ​​model), but has not yet executed the AI ​​service-related computation task on the network device side, the third information may include the input data of the AI ​​service; that is, the input data of the AI ​​service can be understood as the input data of the AI ​​service-related computation task that the network device side needs to execute. When the first network device has already completed part of the AI ​​service-related computation task and obtained intermediate result data, the third information may include the intermediate result data of the AI ​​service-related computation task. When the first network device has completed the AI ​​service-related computation task on the network device side and obtained the final result of the AI ​​service, the third information may include the final result of the AI ​​service; that is, the final result of the AI ​​service is transmitted through AI service data. Additionally, when the first message carries a third request, the third information may also include the AI ​​model related to the AI ​​service.

[0148] The input data for the AI ​​model refers to the AI ​​service data sent by the first terminal before it has been processed. The intermediate result data of the AI ​​service-related computing tasks refers to the data output by the first network device after performing part of the computing tasks, or the data after performing intermediate steps of AI inference that can be divided into different steps.

[0149] In order to ensure a smooth AI service switchover and maintain the continuity of AI services, in one embodiment, the third information is carried in the extension header of the user plane protocol.

[0150] Here, the third information can be carried in the extension header of the user plane protocol and sent to the second network device through the Xn user plane interface. The user plane protocol can be the General Packet Radio Service (GPRS) User Plane Tunneling Protocol for the User Plane (GTP-U); the third information can be understood as AI Service Data. Carrying the third information in the extension header of the user plane protocol can be understood as marking the third information in the user plane protocol through the extension header so that the second network device can identify and process it.

[0151] The extension header may also include: PDU Type, Quick Medium Access Control Protocol (QMP) extension header, Downlink Delay Ind., Uplink Delay Ind., Sequence Number Protection (SNP) extension header, N3 / N9 Interface Delay Ind., New IE Flag, QoS Flow Identifier (QFI), Downlink Sending Time Stamp Repeated, Downlink Received Time Stamp, Uplink Sending Time Stamp (ULSending Time Stamp), Downlink Delay Result, Uplink Delay Result (ULDelay Result), Uplink QoS Flow Identifier Sequence Number (UL QFI Sequence Number), N3 / N9 Interface Delay Result, and AI Service data. The data includes one or more of the following fields: Data, Spare, D1 UL PDCP Delay Result Ind, and Padding. Table 1 shows the location of AI service data in the GTP-U extension header.

[0152] Table 1 GTP-U Extension Headers

[0153]

[0154]

[0155] In order to enable the interaction of AI service-related data or messages between two network devices during AI service switching, in one embodiment, the second data or calculation results related to the AI ​​service are transmitted through the user plane connection, and the request and confirmation messages related to the AI ​​service are transmitted through the control plane connection.

[0156] Here, the second data related to AI services includes the data required to execute AI services, such as AI models, intermediate result data of AI service-related computational tasks, input data of AI services, and final results of AI services; the requests related to AI services include first requests, second requests, and third requests; and the confirmation messages related to AI services include first messages.

[0157] It should be noted that during AI service switching, a connection needs to be established between the control plane and user plane of the AI ​​service signaling and data Xn interface for communication of AI service-related data between the two network devices during the AI ​​service switching process. For example, ... Figure 3 As shown, in the network architecture for AI services, ① represents the connection between the AI ​​service control function and the control plane of the Xn interface, and ② represents the connection between the AI ​​service data forwarding function and the user plane of the Xn interface. The specific network architecture used in actual applications is not specifically limited here.

[0158] Figure 3 The network architecture shown, designed for AI services, adds protocol layer functionalities compared to the 5G network architecture. These functionalities include AI task orchestration, AI task management, a computation execution unit, AI model storage, a forwarding module, and a Computing Data Adaptation Protocol (CDAP). Specifically, AI task orchestration initiates AI service functions and configures computing resources, such as launching AI service functions and containers; AI task management resolves the Quality of Service (QoS) requirements of AI service requests from the first terminal and sends the resolution results to the Radio Resource Control (RRC) unit, AI model storage, computation execution unit, data forwarding, and CDAP module; the computation execution unit performs AI model computations, including typical inference and training services; AI model storage stores AI models of varying accuracy; the forwarding module forwards AI service-related data, specifically forwarding it to the first terminal, cooperating base stations, and the cloud; since AI service-related data needs to be transmitted to the computation unit, a CDAP layer can be added to the protocol stack for parsing AI services. It should be noted that control signaling can be issued by the control node of the AI ​​service, and AI service-related data can be issued by the forwarding module. The format of the signaling and related data forwarded when the AI ​​service switches is compatible with the current communication mechanism.

[0159] Figure 3In the first terminal, the user plane includes: Application (APP) layer, Service Data Adaptation Protocol (SDAP) layer, Packet Data Convergence Protocol (PDCP) layer, Radio Link Control (RLC) layer, Media Access Control (MAC) layer, and Physical Layer (PHY); the control plane of the first terminal includes: Non-Access Stratum (NAS), RRC layer, PDCP layer, RLC layer, MAC layer, and PHY layer. The control plane of network devices includes: RRC layer, PDCP layer, RLC layer, MAC layer, PHY layer, Next Generation Access Protocol (NGAP) layer, Xn Application Protocol layer, Stream Control Transmission Protocol (SCTP) layer, and Ethernet Industrial Protocol (ETH / IP) layer. The user plane of network devices includes: the newly added CDAP layer, SDAP layer, PDCP layer, RLC layer, MAC layer, PHY layer, GTPU layer, and User Datagram Protocol (UDP) / ETH / IP layer. The AMF control plane includes: 5G NAS layer, NGAP layer, SCTP layer, and ETH / IP layer. The UPF user plane includes: GTPU layer, UDP layer, and ETH / IP layer.

[0160] This application also provides a handover method applied to a second network device, which can be understood as a target base station. That is, after network device handover, the base station continues to establish a connection with the first terminal and provides AI services to the first terminal, where the first terminal can be any terminal. Figure 4 As shown, the method includes:

[0161] Step 401: Receive the first request sent by the first network device.

[0162] The first request carries first information, which is used to request the switching of the AI ​​service provided by the first network device to the second network device. The first information represents the relevant requirements of the AI ​​service.

[0163] Here, if the first request includes both a connection switching request and an AI service switching request, preparations are made to perform a connection switching and an AI service switching; if the first request is a connection switching request, preparations are made to perform a connection switching.

[0164] To ensure a smooth AI service switchover and thus guarantee the real-time performance, accuracy, and quality of the AI ​​service, in one embodiment, the first information includes one or more of the following information about the AI ​​service:

[0165] AI models;

[0166] Calculate the container's configuration parameters;

[0167] QoAIS.

[0168] Here, the AI ​​model refers to the AI ​​model required to use or invoke the AI ​​service when executing the AI ​​service. If the first information includes the AI ​​model, the AI ​​model can be invoked during AI service switching to continue providing AI services to the first terminal. If the first information includes the configuration parameters of the computing container, the AI ​​computing container can be started, prepared, or configured based on the configuration parameters of the computing container upon confirmation of AI service switching to invoke the AI ​​model and provide AI services to the first terminal. If the first information includes QoAIS, all QoAIS metrics must be met when providing AI services to the first terminal to ensure the quality of the AI ​​service.

[0169] To enable the first network device to successfully make a handover decision and achieve AI service handover, in one embodiment, before receiving the first request sent by the first network device, the method further includes:

[0170] Receive a second request sent by the first network device, the second request being used to request information related to computing resources;

[0171] Send a second message to the first network device; the second message represents information related to the computing resources of the second network device.

[0172] Here, upon receiving a second request from the first network device, second information may be sent to the first network device based on the current computing power and / or computing resource status; the second information may include parameters indicating idle computing resources, such as the percentage of idle computing resources, and may also include one or more parameters measuring the computing power of the second network device.

[0173] It should be noted that the second network device can also periodically report information about its computing resources to the first network device.

[0174] To enable AI service switching and ensure the continuity of AI services, in one embodiment, after receiving the first request sent by the first network device, the method further includes:

[0175] Based on the first information, decide whether to respond to or reject the first request;

[0176] If the decision is made to respond to the first request, a first message is sent to the first network device, the first message indicating agreement to the handover.

[0177] Here, after receiving the first request from the first network device, admission control or access control is performed based on the first information carried in the first request. Admission control or access control includes connection admission or connection access control, and / or computational admission or computational access control. For computational admission or computational access control, whether to agree to the AI ​​service switch can be determined by whether the QoAIS metrics included in the first information are supported. If the QoAIS metrics included in the first information are supported, the first request is responded to, and a first message is sent to the first network device. If the QoAIS metrics included in the first information are not supported, the first request is rejected, and a second message is sent to the first network device, indicating that the switch is not agreed to.

[0178] It should be noted that if the first information does not include an AI model, the first message may carry a third request to request the acquisition or sharing of the AI ​​model; that is, if the first information does not include an AI model and the second network device can meet the QoAIS indicators, the second network device is considered to support the QoAIS indicators included in the first information and support providing the AI ​​services required by the first terminal.

[0179] It should be noted that when the first request includes both a connection switching request and an AI service switching request, a response can be given to both the connection switching request and the AI ​​service switching request; that is, the first message can indicate confirmation of connection switching and confirmation of AI service switching; when the first request is a connection switching request, the first message can indicate confirmation of connection switching.

[0180] In order to continue providing AI services to the first terminal and ensure the continuity of AI services, in one embodiment, the method further includes:

[0181] If the second network device meets the relevant requirements of the AI ​​service represented by the first information, a switching preparation is performed based on the configuration parameters of the computing container related to the AI ​​service.

[0182] Here, if the second network device satisfies the QoAIS included in the first information, the AI ​​computing container is configured based on the configuration parameters of the computing container included in the first information, and the AI ​​computing container is started or prepared so as to ensure the continuity and real-time performance of the AI ​​service after the network device is switched.

[0183] To ensure the continuity and real-time nature of AI services, in one embodiment, the method further includes:

[0184] The system receives third information sent by the first network device, the third information representing information used to perform the AI ​​service or the calculation result of the AI ​​service.

[0185] Here, when a first message is sent to the first network device, a third message is received from the first network device; when the third message is received from the first network device, AI services can continue to be provided to the first terminal based on the third message after the network device is switched, ensuring the continuity and real-time nature of the AI ​​service switchover.

[0186] It should be noted that it can also receive and store the fourth information sent by the first network device, and continue to transmit information to the first terminal based on the fourth information after the connection switch, so as to ensure the continuity and real-time nature of the connection switch.

[0187] To ensure the continuity and real-time nature of the AI ​​service, in one embodiment, after receiving the third information sent by the first network device, the method further includes:

[0188] Store the third information; and / or

[0189] Based on the third information, perform the computational tasks related to the AI ​​service that the first network device has not completed.

[0190] Here, when the third information indicates that the AI ​​service-related computational task has been completed (i.e., the third information includes the final result of the AI ​​service), the third information is stored. When the third information indicates that the AI ​​service-related computational task has not yet been completed or that some part remains to be completed (i.e., the third information includes the input data of the AI ​​service or the intermediate result data of the AI ​​service-related computational task), the AI ​​service-related computational task that the first network device has not completed is executed based on the third information. When the first network device provides multiple AI services to the first terminal, it can receive third information corresponding to different AI services. These different AI services may include a first AI service whose related computational task has been completed and a second AI service whose related computational task has not yet been completed. The second network device can store the third information corresponding to the first AI service and execute the remaining computational task based on the third information of the second AI service.

[0191] In order to continue providing AI services to the first terminal and ensure the continuity of AI services, in one embodiment, the third information includes one or more of the following:

[0192] The AI ​​model related to the AI ​​service;

[0193] The input data for the AI ​​service;

[0194] Intermediate result data of the computational tasks related to the AI ​​service;

[0195] The final result of the AI ​​service.

[0196] Here, if the third information includes an AI model related to the AI ​​service, the AI ​​model is placed into a computing container to continue providing AI services to the first terminal. If the third information includes input data for the AI ​​service, the input data is input into the AI ​​model to execute AI service-related computing tasks and provide AI services to the first terminal. If the third information includes intermediate result data of the AI ​​service-related computing tasks, the intermediate result data is input into the corresponding intermediate node in the AI ​​model to continue executing the remaining AI service-related computing tasks. If the third information includes the final result of the AI ​​service, it is indicated that the AI ​​service has been completed, and the final result of the AI ​​service is stored so that it can be sent to the first terminal after the switch is completed.

[0197] To save on signaling overhead and to ensure compatibility with current communication formats, in one embodiment, the third information is carried in the extension header of the user plane protocol.

[0198] Here, third-party information can be obtained by identifying and processing the extension header of the user plane protocol.

[0199] The following section provides a more detailed description of this application with reference to application examples.

[0200] like Figure 5 As shown, the method for switching network devices includes the following steps:

[0201] Step 1: The first network device sends the measurement configuration to the first terminal.

[0202] Here, the triggering events for AI service switching include connection switching triggered when the first terminal is mobile, and the first terminal's currently connected network device lacking sufficient computing resources, requiring it to request computing resources from a nearby network device. This application embodiment addresses the first triggering event. Simultaneously with the connection switching, the AI ​​service switching is triggered, and the first network device sends measurement configurations to the first terminal.

[0203] Step 2: The first terminal reports the measurement results to the first network device.

[0204] Here, upon receiving the measurement configuration sent by the first network device, the first terminal measures parameters such as signal strength, signal quality, and interference level based on the measurement configuration, obtains the measurement results, and reports the measurement results to the first network device.

[0205] Step 3: The first network device sends a second request to the second network device. The second request is used to request information related to computing resources.

[0206] Here, upon receiving the measurement results, the first network device sends a second request to the second network device based on the data in the measurement results.

[0207] Step 4: The second network device sends second information to the first network device. The second information represents information related to the computing resources of the second network device.

[0208] Here, upon receiving a second request from the first network device, the second network device collects relevant information about its computing resources, including the percentage of idle computing resources, and may also include one or more parameters that measure the computing power of the second network device, to obtain second information; and sends the second information to the first network device.

[0209] Step 5: The first network device makes a handover decision.

[0210] Here, upon receiving the measurement results and the second information, the first network device makes connection switching decisions and AI service switching decisions. The specific decision-making method is as follows:

[0211] If the signal quality of the second network device meets the handover requirements, and one or more of the following conditions are met, the first network device decides to perform connection handover and AI service handover:

[0212] The second network device supports the AI ​​service;

[0213] The first network device receives first data related to the AI ​​service sent by the first terminal;

[0214] The first network device has not yet started executing the computational tasks related to the AI ​​service;

[0215] The first network device has not completed the computational tasks related to the AI ​​service;

[0216] The remaining computational load of the AI ​​service in the first network device is greater than or equal to the first threshold.

[0217] It should be noted that, in the case where the first network device periodically acquires and stores information related to the computing resources of the second network device, the first network device can make a switching decision directly based on the latest stored information related to the computing resources of the second network device and the terminal measurement results after receiving the measurement results from the terminal, without needing to send a second request to the second network device based on the terminal measurement results. That is, step 5 can be executed directly after step 2.

[0218] Step 6: The first network device sends a first request to the second network device.

[0219] Here, when the first network device decides to switch the connection and switch the AI ​​service, it sends a first request to the second network device to request the connection switch and to switch the AI ​​service provided by the first network device to the second network device.

[0220] Step 7: The second network device decides whether to respond to or reject the first request based on the first information.

[0221] Here, when the second network device receives the first request, it determines whether it supports the QoAIS metrics included in the first information based on the first information carried in the first request. If the second network device supports the QoAIS metrics included in the first information, it decides to respond to the first request; if it does not support the QoAIS metrics included in the first information, it decides to reject the first request.

[0222] Step 8: The second network device sends the first message to the first network device.

[0223] Here, the second network device sends a first message to the first network device after deciding to respond to the first request, that is, to accept the access of the first terminal and to provide AI services to the first terminal.

[0224] Step 9: The first network device and the first terminal perform synchronization configuration and RRC reconfiguration.

[0225] Here, upon receiving the first message, the first network device prepares to perform connection switching and AI service switching, and synchronizes the switching message to the first terminal, that is, sends a switching command and an RRC reconfiguration message to the first terminal; if there is a computing bearer between the first terminal and the first network device, the first network device also needs to send a computing bearer reconfiguration message to the first terminal.

[0226] Step 10: The first network device sends third information to the second network device. The third information represents the information used to perform the AI ​​service or the calculation result of the AI ​​service.

[0227] Here, upon receiving the first message, the first network device obtains information or calculation results of the AI ​​service executed by the first terminal, obtains the third information, and sends the third information to the second network device. Steps 10 and 9 can be performed simultaneously.

[0228] Step 11: The first network device sends an SN status synchronization message to the second network device.

[0229] Here, the first network device sends a sequence number status transmission message to the second network device.

[0230] Step 12: The second network device stores the third information and / or performs computational tasks related to the AI ​​service that the first network device has not completed based on the third information.

[0231] Here, upon receiving the third information, the second network device determines the execution status of the AI ​​service of the first terminal based on the third information, selects to store the third information, and / or executes the computational tasks related to the AI ​​service that the first network device has not completed based on the third information.

[0232] Step 13: The first terminal performs synchronization with the second network device.

[0233] Here, when the first terminal receives the handover command sent by the first network device, it performs synchronization with the second network device and accesses the target cell in the second network device; the synchronization includes frequency synchronization, time synchronization, frame synchronization and downlink synchronization.

[0234] Step 14: The second network device responds with the uplink resource allocation instruction and performs terminal timing.

[0235] Here, the second network device responds with an uplink resource allocation instruction to indicate the uplink resources allocated to the first terminal, and sends a timing advance command to the first terminal for timing correction.

[0236] Step 15: The first terminal sends an RRC reconfiguration complete message to the second network device.

[0237] Here, when the first terminal successfully accesses the target cell in the second network device, it sends an RRC reconfiguration complete message to the second network device so that the second network device can confirm the successful handover and send data to the first terminal.

[0238] To implement the method on the first network device side of this application embodiment, this application embodiment also provides a switching device, disposed on the first network device, such as... Figure 6 As shown, the device includes:

[0239] The first sending unit 601 is used to send a first request to the second network device; wherein the first request carries first information, which is used to request to switch the artificial intelligence (AI) service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0240] In one embodiment, the first information includes one or more of the following information from the AI ​​service:

[0241] AI models;

[0242] Calculate the container's configuration parameters;

[0243] AI Service Quality (QoAIS)

[0244] In one embodiment, the first sending unit 601 is specifically configured to send a first request to the second network device when the signal quality of the second network device meets the handover requirements and one or more of the following conditions are met:

[0245] The second network device supports the AI ​​service;

[0246] The first network device receives first data related to the AI ​​service sent by the first terminal;

[0247] The first network device has not yet started executing the computational tasks related to the AI ​​service;

[0248] The first network device has not completed the computational tasks related to the AI ​​service;

[0249] The remaining computational load of the AI ​​service in the first network device is greater than or equal to the first threshold.

[0250] In one embodiment, the device further includes:

[0251] The second sending unit is used to send a second request to the second network device before sending the first request to the second network device. The second request is used to request information related to computing resources.

[0252] The second receiving unit is configured to receive second information sent by the second network device; the second information represents information related to the computing resources of the second network device.

[0253] In one embodiment, the transpose further includes:

[0254] The switching unit is used to determine, based on the second information, whether the second network device supports or does not support the AI ​​service.

[0255] In one embodiment, the device further includes:

[0256] The third receiving unit is configured to receive a first message sent by the second network device, wherein the first message indicates agreement to switchover.

[0257] The third sending unit is used to send third information to the second network device, the third information representing information for performing the AI ​​service or the calculation result of the AI ​​service.

[0258] In one embodiment, the third information includes one or more of the following:

[0259] The AI ​​model related to the AI ​​service;

[0260] The input data for the AI ​​service;

[0261] Intermediate result data of the computational tasks related to the AI ​​service;

[0262] The final result of the AI ​​service.

[0263] In one embodiment, the third information is carried in the extension header of the user plane protocol.

[0264] In one embodiment, the second data or calculation results related to the AI ​​service are transmitted via a user plane connection, and the requests and confirmation messages related to the AI ​​service are transmitted via a control plane connection.

[0265] In practical applications, the first transmitting unit 601, the second transmitting unit, the second receiving unit, the third receiving unit, and the third transmitting unit can be implemented by a processor in the switching device combined with a communication interface; the switching unit can be implemented by a processor in the switching device.

[0266] To implement the method on the second network device side of the embodiments of this application, the embodiments of this application also provide a switching device, which is disposed on the second network device, such as... Figure 7 As shown, the device includes:

[0267] The first receiving unit 701 is configured to receive a first request sent by a first network device; wherein the first request carries first information, which is used to request to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0268] In one embodiment, the first information includes one or more of the following information from the AI ​​service:

[0269] AI models;

[0270] Calculate the container's configuration parameters;

[0271] QoAIS.

[0272] In one embodiment, the device further includes:

[0273] The fourth receiving unit is configured to receive a second request sent by the first network device before receiving the first request sent by the first network device, wherein the second request is used to request information related to computing resources;

[0274] The fourth sending unit is used to send second information to the first network device; the second information represents information related to the computing resources of the second network device.

[0275] In one embodiment, the device further includes:

[0276] The decision unit is configured to, after receiving a first request sent by a first network device, decide whether to respond to or reject the first request based on the first information;

[0277] The fifth sending unit is configured to send a first message to the first network device, indicating agreement to switch, if it decides to respond to the first request.

[0278] In one embodiment, the device further includes:

[0279] The fifth receiving unit is used to receive third information sent by the first network device, wherein the third information represents information for performing the AI ​​service or the calculation result of the AI ​​service.

[0280] In one embodiment, the device further includes:

[0281] The preparation unit is used to perform switching preparation based on the configuration parameters of the computing container related to the AI ​​service, provided that the second network device meets the relevant requirements of the AI ​​service represented by the first information.

[0282] In one embodiment, the device further includes:

[0283] A storage unit is configured to store the third information after receiving the third information sent by the first network device; and / or

[0284] An execution unit is configured to perform computational tasks related to the AI ​​service that the first network device has not completed, based on the third information.

[0285] In one embodiment, the third information includes one or more of the following:

[0286] The AI ​​model related to the AI ​​service;

[0287] The input data for the AI ​​service;

[0288] Intermediate result data of the computational tasks related to the AI ​​service;

[0289] The final result of the AI ​​service.

[0290] In one embodiment, the third information is carried in the extension header of the user plane protocol.

[0291] In practical applications, the first receiving unit 701, the fourth receiving unit, the fourth sending unit, the fifth sending unit, and the fifth receiving unit can be implemented by a processor in the switching device combined with a communication interface, and the decision unit, the preparation unit, the storage unit, and the execution unit can be implemented by a processor in the switching device.

[0292] It should be noted that the switching device provided in the above embodiments is only illustrated by the division of the above program modules. In practical applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the processing described above. In addition, the switching device and the switching method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0293] Based on the hardware implementation of the above program modules, and in order to implement the method on the first network device side of the embodiments of this application, the embodiments of this application also provide a first network device, such as... Figure 8 As shown, the first network device 800 includes:

[0294] The first communication interface 801 can exchange information with other network nodes.

[0295] The first processor 802 is connected to the first communication interface 801 to enable information interaction with other network nodes and to execute the methods provided by one or more technical solutions on the first network device side when running a computer program.

[0296] The first memory 803 is used to store computer programs that can run on the first processor 802.

[0297] Specifically, the first communication interface 801 is used to send a first request to the second network device; wherein, the first request carries first information, which is used to request to switch the artificial intelligence (AI) service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0298] In one embodiment, the first information includes one or more of the following information from the AI ​​service:

[0299] AI models;

[0300] Calculate the container's configuration parameters;

[0301] AI Service Quality (QoAIS)

[0302] In one embodiment, the first communication interface 801 is specifically used to send a first request to the second network device when the signal quality of the second network device meets the switching requirements and one or more of the following conditions are met:

[0303] The second network device supports the AI ​​service;

[0304] The first network device receives first data related to the AI ​​service sent by the first terminal;

[0305] The first network device has not yet started executing the computational tasks related to the AI ​​service;

[0306] The first network device has not completed the computational tasks related to the AI ​​service;

[0307] The remaining computational load of the AI ​​service in the first network device is greater than or equal to the first threshold.

[0308] In one embodiment, the first communication interface 801 is further configured to send a second request to the second network device before sending the first request to the second network device, the second request being used to request information related to computing resources;

[0309] The system receives second information sent by the second network device; the second information represents information related to the computing resources of the second network device.

[0310] In one embodiment, the first processor 802 is configured to determine, based on the second information, whether the second network device supports or does not support the AI ​​service.

[0311] In one embodiment, the first communication interface 801 is further configured to receive a first message sent by the second network device, the first message indicating agreement to switch;

[0312] Send a third message to the second network device, the third message representing information for performing the AI ​​service or the calculation result of the AI ​​service.

[0313] In one embodiment, the third information includes one or more of the following:

[0314] The AI ​​model related to the AI ​​service;

[0315] The input data for the AI ​​service;

[0316] Intermediate result data of the computational tasks related to the AI ​​service;

[0317] The final result of the AI ​​service.

[0318] In one embodiment, the third information is carried in the extension header of the user plane protocol.

[0319] In one embodiment, the second data or calculation results related to the AI ​​service are transmitted via a user plane connection, and the requests and confirmation messages related to the AI ​​service are transmitted via a control plane connection.

[0320] It should be noted that the specific processing procedures of the first processor 802 and the first communication interface 801 can be understood by referring to the above method. Of course, in practical applications, the various components in the first network device 800 are coupled together through the bus system 804. It can be understood that the bus system 804 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 804 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, in... Figure 8 The general labeled all buses as Bus System 804.

[0321] The first memory 803 in this embodiment is used to store various types of data to support the operation of the first network device 800. Examples of such data include any computer program used to operate on the first network device 800.

[0322] The methods disclosed in the embodiments of this application can be applied to the first processor 802, or implemented by the first processor 802. The first processor 802 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware or by instructions in the form of software in the first processor 802. The first processor 802 may be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The first processor 802 can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly reflected as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium, which is located in the first memory 803. The first processor 802 reads the information in the first memory 803 and completes the steps of the aforementioned method in combination with its hardware.

[0323] In an exemplary embodiment, the first network device 800 may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components to perform the aforementioned method.

[0324] Based on the hardware implementation of the above program modules, and in order to implement the method on the second network device side of the embodiments of this application, the embodiments of this application also provide a second network device, such as... Figure 9 As shown, the second network device 900 includes:

[0325] The first communication interface 901 can exchange information with other network nodes.

[0326] The first processor 902 is connected to the first communication interface 901 to enable information interaction with other network nodes and to execute the methods provided by one or more technical solutions on the second network device side when running computer programs.

[0327] The first memory 903 is used to store computer programs that can run on the first processor 902.

[0328] Specifically, the first communication interface 901 is used to receive a first request sent by the first network device; wherein the first request carries first information, which is used to request to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

[0329] In one embodiment, the first information includes one or more of the following information from the AI ​​service:

[0330] AI models;

[0331] Calculate the container's configuration parameters;

[0332] QoAIS.

[0333] In one embodiment, the second communication interface 901 is further configured to receive a second request sent by the first network device before receiving the first request sent by the first network device, the second request being used to request information related to computing resources;

[0334] Send a second message to the first network device; the second message represents information related to the computing resources of the second network device.

[0335] In one embodiment, the second processor 902 is further configured to, after receiving a first request sent by the first network device, decide whether to respond to or reject the first request based on the first information;

[0336] The second communication interface 901 is further configured to send a first message to the first network device, wherein the first message indicates agreement to switch, if the decision is made to respond to the first request.

[0337] In one embodiment, the second communication interface 901 is further configured to receive third information sent by the first network device, the third information representing information for performing the AI ​​service or the calculation result of the AI ​​service.

[0338] In one embodiment, the second processor 902 is further configured to perform switching preparation based on the configuration parameters of the computing container related to the AI ​​service, provided that the second network device meets the relevant requirements of the AI ​​service represented by the first information.

[0339] In one embodiment, the second processor 902 is further configured to store the third information after receiving the third information sent by the first network device; and / or

[0340] Based on the third information, perform the computational tasks related to the AI ​​service that the first network device has not completed.

[0341] In one embodiment, the third information includes one or more of the following:

[0342] The AI ​​model related to the AI ​​service;

[0343] The input data for the AI ​​service;

[0344] Intermediate result data of the computational tasks related to the AI ​​service;

[0345] The final result of the AI ​​service.

[0346] In one embodiment, the third information is carried in the extension header of the user plane protocol.

[0347] It should be noted that the specific processing procedures of the second processor 902 and the second communication interface 901 can be understood by referring to the above method. Of course, in practical applications, the various components in the second network device 900 are coupled together through the bus system 904. It can be understood that the bus system 904 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 904 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, in... Figure 9 The general designated all buses as Bus System 904.

[0348] The second memory 903 in this embodiment is used to store various types of data to support the operation of the second network device 900. Examples of such data include any computer program used to operate on the second network device 900.

[0349] The methods disclosed in the embodiments of this application can be applied to, or implemented by, the second processor 902. The second processor 902 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by the integrated logic circuitry of the hardware or by instructions in the form of software within the second processor 902. The second processor 902 may be a general-purpose processor, a DSP, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The second processor 902 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium, specifically a second memory 903. The second processor 902 reads information from the second memory 903 and, in conjunction with its hardware, completes the steps of the aforementioned method.

[0350] In an exemplary embodiment, the second network device 900 may be implemented by one or more ASICs, DSPs, PLDs, CPLDs, FPGAs, general-purpose processors, controllers, MCUs, microprocessors, or other electronic components to perform the aforementioned method.

[0351] It is understood that the memories (first memory 803 and second memory 903) in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile memory and non-volatile memory. Specifically, non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).The memories described in the embodiments of this application are intended to include, but are not limited to, these and any other suitable types of memories.

[0352] In an exemplary embodiment, this application also provides a storage medium, namely a computer storage medium, specifically a computer-readable storage medium, such as a first memory 803 storing a computer program, which can be executed by a first processor 802 of a first network device 800 to complete the steps described in the aforementioned first network device-side method. Another example is a second memory 903 storing a computer program, which can be executed by a second processor 902 of a second network device 900 to complete the steps described in the aforementioned second network device-side method. The computer-readable storage medium can be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM.

[0353] For example, embodiments of this application also provide a computer program product, including a computer program that can be executed by a first processor 802 of a first network device 800 and a second processor 902 of a second network device 900 to perform the steps described in any of the foregoing methods.

[0354] It should be noted that terms such as "first" and "second" are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence; "multiple" refers to two or more items. The term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, the technical solutions described in the embodiments of this application can be arbitrarily combined without conflict. The above descriptions are merely preferred embodiments of this application and are not intended to limit the scope of protection of this application.

Claims

1. A switching method, characterized in that, Applied to a first network device, the method includes: Send a first request to a second network device; wherein the first request carries first information for requesting to switch the artificial intelligence (AI) service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

2. The method according to claim 1, characterized in that, The first information includes one or more of the following information from the AI ​​service: AI models; Calculate the container's configuration parameters; AI Service Quality (QoAIS) 3. The method according to claim 1, characterized in that, Sending the first request to the second network device includes: A first request is sent to the second network device if the signal quality of the second network device meets the handover requirements and one or more of the following conditions are met: The second network device supports the AI ​​service; The first network device receives first data related to the AI ​​service sent by the first terminal; The first network device has not yet started executing the computational tasks related to the AI ​​service; The first network device has not completed the computational tasks related to the AI ​​service; The remaining computational load of the AI ​​service in the first network device is greater than or equal to the first threshold.

4. The method according to any one of claims 1 to 3, characterized in that, Before sending the first request to the second network device, the method further includes: Send a second request to the second network device, the second request being used to request information related to computing resources; The system receives second information sent by the second network device; the second information represents information related to the computing resources of the second network device.

5. The method according to claim 4, characterized in that, The method further includes: Based on the second information, it is determined whether the second network device supports or does not support the AI ​​service.

6. The method according to any one of claims 1 to 3, characterized in that, The method further includes: Receive a first message sent by the second network device, the first message indicating agreement to switch; Send a third message to the second network device, the third message representing information for performing the AI ​​service or the calculation result of the AI ​​service.

7. The method according to claim 6, characterized in that, The third information includes one or more of the following: The AI ​​model related to the AI ​​service; The input data for the AI ​​service; Intermediate result data of the computational tasks related to the AI ​​service; The final result of the AI ​​service.

8. The method according to claim 6, characterized in that, The third information is carried in the extension header of the user plane protocol.

9. The method according to any one of claims 1 to 3, 5, and 7 to 8, characterized in that, The second data or calculation results related to the AI ​​service are transmitted through the user plane connection, and the requests and confirmation messages related to the AI ​​service are transmitted through the control plane connection.

10. A switching method, characterized in that, Applied to a second network device, the method includes: Receive a first request sent by a first network device; wherein the first request carries first information for requesting to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

11. The method according to claim 10, characterized in that, The first information includes one or more of the following information from the AI ​​service: AI models; Calculate the container's configuration parameters; QoAIS.

12. The method according to claim 10 or 11, characterized in that, Before receiving the first request sent by the first network device, the method further includes: Receive a second request sent by the first network device, the second request being used to request information related to computing resources; Send a second message to the first network device; the second message represents information related to the computing resources of the second network device.

13. The method according to claim 10 or 11, characterized in that, After receiving the first request sent by the first network device, the method further includes: Based on the first information, decide whether to respond to or reject the first request; If the decision is made to respond to the first request, a first message is sent to the first network device, the first message indicating agreement to the handover.

14. The method according to claim 13, characterized in that, The method further includes: The system receives third information sent by the first network device, the third information representing information used to perform the AI ​​service or the calculation result of the AI ​​service.

15. The method according to claim 11, characterized in that, The method further includes: If the second network device meets the relevant requirements of the AI ​​service represented by the first information, a switching preparation is performed based on the configuration parameters of the computing container related to the AI ​​service.

16. The method according to claim 14, characterized in that, After receiving the third information sent by the first network device, the method further includes: Store the third information; and / or Based on the third information, perform the computational tasks related to the AI ​​service that the first network device has not completed.

17. The method according to claim 14, characterized in that, The third information includes one or more of the following: The AI ​​model related to the AI ​​service; The input data for the AI ​​service; Intermediate result data of the computational tasks related to the AI ​​service; The final result of the AI ​​service.

18. The method according to claim 14, characterized in that, The third information is carried in the extension header of the user plane protocol.

19. A switching device, characterized in that, include: The first sending unit is configured to send a first request to the second network device; wherein the first request carries first information, which is used to request the switching of the artificial intelligence (AI) service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

20. A switching device, characterized in that, include: The first receiving unit is configured to receive a first request sent by a first network device; wherein the first request carries first information, which is used to request switching the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

21. A first network device, characterized in that, include: A first processor and a first communication interface; wherein the first communication interface is used to send a first request to a second network device; wherein the first request carries first information for requesting to switch the artificial intelligence (AI) service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

22. A second network device, characterized in that, include: A second processor and a second communication interface; wherein... The second communication interface is used to receive a first request sent by the first network device; wherein the first request carries first information, which is used to request to switch the AI ​​service provided by the first network device to the second network device, and the first information represents the relevant requirements of the AI ​​service.

23. A network device, characterized in that, This includes a processor and memory for storing computer programs that can run on the processor. When the processor is used to run the computer program, it performs the steps of the method according to any one of claims 1 to 9, or performs the steps of the method according to any one of claims 10 to 18.

24. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9, or the steps of the method according to any one of claims 10 to 18.

25. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 18.

Images