Search

CN-122028112-A - Computing power service request processing system, method, device, computer equipment and medium

CN122028112ACN 122028112 ACN122028112 ACN 122028112ACN-122028112-A

Abstract

The application relates to a system, a method, a device, computer equipment and a medium for processing a power calculation service request. The method comprises the steps that under the condition that a target user initiates a calculation service request, a terminal obtains a calculation indicator carrying user intention of the target user and uploads the calculation indicator to a cloud end, the cloud end forwards the calculation indicator to an edge side, the terminal is instructed to interactively process the calculation service request with the edge side according to the calculation indicator, geographic position information of the terminal and SLA guarantee requirements of the target user, the edge side obtains a KQI index matched with the calculation service request according to the current network condition of a network to which the edge side belongs and the calculation indicator, and under the condition that the KQI index does not meet SLA guarantee requirements, the request indicator of the target user is generated and sent to the cloud end, and finally the cloud end cooperatively processes the calculation service request with the terminal and provides calculation support for the terminal according to the user intention, the calculation indicator and the request indicator. The method improves the processing efficiency of the calculation service request.

Inventors

  • WU HAIBO
  • ZHAO JING

Assignees

  • 中国电信股份有限公司技术创新中心
  • 中国电信股份有限公司

Dates

Publication Date
20260512
Application Date
20260107

Claims (20)

  1. 1. The system for processing the computing power business request is characterized by comprising a terminal, an edge side and a cloud end; The terminal is used for acquiring a terminal computing power indicator and uploading the terminal computing power indicator to the cloud end under the condition that a target user initiates a computing power service request, wherein the terminal computing power indicator carries the user intention of the target user; The cloud end is used for forwarding the terminal computing power indicator to the edge side and indicating the terminal and the edge side to interactively process the computing power service request according to the terminal computing power indicator, the geographic position information of the terminal and the SLA guarantee requirement of the target user; the edge side is used for acquiring a KQI index matched with the power calculation service request according to the current network condition of the network to which the edge side belongs and the terminal power calculation indicator; The cloud end is used for cooperatively processing the computing power service request with the terminal according to the user intention, the terminal computing power indicator and the computing power service request indicator and providing computing power support of a cloud end large model for the terminal.
  2. 2. The system of claim 1, wherein the system further comprises: And the cloud end is further used for determining an anchoring UPF corresponding to the terminal according to the geographic position information of the terminal and the SLA guarantee requirement under the indication of the terminal power calculation symbol so as to indicate the terminal to interactively process the power calculation service request based on the anchoring UPF and the edge side.
  3. 3. The system of claim 2, wherein the system further comprises: The cloud end is further used for obtaining a KQI guarantee required by the target user according to the calculation service request indicator, determining a new anchoring UPF from candidate UPFs according to real-time cell position information of the candidate UPFs when the KQI guarantee representation needs to be replaced by the anchoring UPF, processing the calculation service request in cooperation with the terminal according to the new anchoring UPF and the user intention, and providing calculation support of a cloud end large model for the terminal; The cloud end is further configured to send a user session update suggestion to the PCF to enable the PCF to establish an association with the new anchor UPF.
  4. 4. A system according to claim 3, wherein the system further comprises: The cloud end is further configured to subscribe, through the NWDAF, to the new anchor UPF for reporting experience data of the terminal, where subscription information carries a user identifier, so that the new anchor UPF uploads experience data of a target user corresponding to the user identifier to the NWDAF, and the NWDAF predicts a network support service situation according to the experience data, generates an initial SLA guarantee policy suggestion based on the predicted network support service situation and a requirement of the terminal, and returns the initial SLA guarantee policy suggestion to the cloud end; The cloud end is further used for receiving the initial SLA guarantee strategy suggestion, planning a scheme according to the initial SLA guarantee strategy suggestion and the user intention to obtain a corresponding execution scheme, and processing the computing power service request according to the execution scheme.
  5. 5. A system according to claim 3, wherein the system further comprises: The cloud end is further configured to perform uplink/downlink CSI channel switching at a physical layer, perform bandwidth adaptation at a network layer, and establish a QoS dedicated policy at an application layer, perform service requirement decomposition, calculation and coordination matched with service characteristics of the power computing service request to generate an SLA guarantee policy suggestion, send the SLA guarantee policy suggestion and a policy conflict priority to the PCF, and the PCF is configured to generate a corresponding SLA guarantee rule according to the SLA guarantee policy suggestion, the policy conflict priority and location information of the terminal, and send the corresponding SLA guarantee rule to the SMF, so that the SMF establishes dedicated load according to the SLA guarantee rule, and if the dedicated load suggestion is successful, feedback a dedicated load establishment success indication to the NWDAF through the cloud end.
  6. 6. The system of claim 4, wherein the system further comprises: The cloud end is further configured to receive the updated initial SLA guarantee policy suggestion sent by the NWDAF under the condition that the SLA guarantee policy suggestion cannot meet the SLA guarantee requirement, perform scheme planning according to the updated initial SLA guarantee policy suggestion and the user intention to obtain a corresponding new execution scheme, process the calculation service request according to the new execution scheme, receive new experience data of the target user reported by the new anchor UPF by the updated initial SLA guarantee policy suggestion by the NWDAF, predict a network support service condition according to the new experience data, obtain a new predicted network support service condition, and update the initial SLA guarantee policy suggestion based on the new predicted network support service condition and the requirements of the terminal.
  7. 7. The system of claim 1, wherein the system further comprises: The edge side is further configured to interact with the terminal to process the power calculation service request when the KQI index meets the SLA guarantee requirement, where the KQI index includes at least one index of a blocking number, a blocking duration, a code rate, a resolution, a response delay, an establishment delay, a jitter, and an error rate.
  8. 8. The system of claim 1, wherein the system further comprises: The terminal is further used for acquiring communication data, user behavior data, personalized preference data, terminal computing power information, load information and link quality information of the terminal, session scene information of the terminal is acquired, generating the terminal computing power indicator based on the session scene information, the terminal computing power information, the load information, the link quality information and the computing power service request, carrying out local processing on the communication data, the user behavior data and the personalized preference data to acquire the user intention, and the local processing comprises personal privacy data protection encryption, semantic query and intention recognition.
  9. 9. The system of claim 8, wherein the system further comprises: The terminal is further used for filtering malicious sensitive content in the communication data, the user behavior data and the personalized preference data to obtain cleaned communication data, user behavior data and personalized preference data, and performing local data processing on the cleaned communication data, the user behavior data and the personalized preference data to obtain the user intention.
  10. 10. A method for processing a computing power service request, which is applied to a cloud, the method comprising: the terminal computing power indicator uploaded by the terminal is received, wherein the terminal computing power indicator is obtained after a target user initiates a computing power service request by the terminal; forwarding the terminal calculation indicator to an edge side, and indicating the terminal to interact with the edge side to process the calculation service request according to the terminal calculation indicator, the geographic position information of the terminal and the SLA guarantee requirement of the target user; Receiving an account service request indicator of the target user sent by the edge side, wherein the account service request indicator is generated by the edge side under the condition that the KQI index does not meet the SLA guarantee requirement according to the current network condition of the network to which the edge side belongs and the terminal account indicator; And according to the user intention, the terminal computing power indicator and the computing power service request indicator, cooperatively processing the computing power service request with the terminal and providing computing power support of a cloud large model for the terminal.
  11. 11. The method according to claim 10, wherein the instructing the terminal to interact with the edge side to process the computing power service request according to the terminal computing power indicator, the geographic location information of the terminal, and the SLA guarantee requirement of the target user includes: and under the indication of the terminal power calculation symbol, determining an anchoring UPF corresponding to the terminal according to the geographic position information of the terminal and the SLA guarantee requirement, so as to indicate the terminal to interactively process the power calculation service request based on the anchoring UPF and the edge side.
  12. 12. The method of claim 11, wherein the processing the computing power service request in conjunction with the terminal according to the user intent, the terminal computing power indicator, and the computing power service request indicator comprises: Obtaining KQI guarantee required by the target user according to the power calculation service request indicator; Under the condition that the KQI guarantee characterization needs to be replaced by an anchor UPF, determining a new anchor UPF from candidate UPFs according to the real-time cell position information of the candidate UPFs, and cooperatively processing the calculation service request with the terminal and providing calculation support of a cloud large model for the terminal according to the new anchor UPF and the user intention; Under the condition that the KQI guarantee characterization does not need to be replaced by an anchor UPF, taking the anchor UPF as a new anchor UPF, and subscribing NWDAF with SLA guarantee strategy suggestion based on the terminal calculation indicator and the calculation service request indicator; User session update suggestions are sent to the PCF to associate the PCF with the new anchor UPF.
  13. 13. The method of claim 12, wherein the processing the computing power service request in conjunction with the terminal according to the new anchor UPF and the user intent comprises: Subscribing experience data of the terminal to the new anchoring UPF through the NWDAF, reporting the experience data, wherein subscription information carries a user identifier, so that the new anchoring UPF uploads the experience data of a target user corresponding to the user identifier to the NWDAF, the NWDAF predicts the network support service condition according to the experience data, and generates an initial SLA guarantee strategy suggestion and returns to the cloud based on the predicted network support service condition and the requirement of the terminal; receiving the initial SLA guarantee strategy suggestion, and planning a scheme according to the initial SLA guarantee strategy suggestion and the user intention to obtain a corresponding execution scheme; and processing the power calculation service request according to the execution scheme.
  14. 14. The method of claim 12, wherein said processing said computing power service request according to said execution scheme comprises: Executing uplink/downlink CSI channel switching at a physical layer, performing bandwidth self-adaption at a network layer, establishing a QoS special-load strategy at an application layer, executing service demand decomposition, calculation and coordination matched with service characteristics of the computing service request, and generating an SLA guarantee strategy suggestion; The PCF is used for generating corresponding SLA guarantee rules according to the SLA guarantee policy suggestions, the policy conflict priorities and the position information of the terminal and sending the corresponding SLA guarantee rules to the SMF so that the SMF establishes a special load according to the SLA guarantee rules, and feeds back a special load establishment success indication to NWDAF through the cloud under the condition that the special load suggestions are successful.
  15. 15. The method of claim 13, wherein the method further comprises: Receiving updated initial SLA guarantee strategy advice sent by NWDAF under the condition that the SLA guarantee strategy advice cannot meet the SLA guarantee requirement, receiving new experience data of the target user reported by the new anchor UPF by NWDAF, predicting network support service conditions according to the new experience data to obtain new predicted network support service conditions, and updating the initial SLA guarantee strategy advice based on the new predicted network support service conditions and the requirements of the terminal to obtain the new predicted network support service conditions; Carrying out scheme planning according to the updated initial SLA guarantee strategy proposal and the user intention to obtain a corresponding new execution scheme; and processing the power calculation service request according to the new execution scheme.
  16. 16. A method for processing a computing power service request, which is applied to an edge side, the method comprising: the terminal computing power indicator forwarded by the cloud is received, the terminal computing power indicator is obtained by a terminal under the condition that a target user initiates a computing power service request and is sent to the cloud, and the terminal computing power indicator carries the user intention of the target user; After the cloud side indicates the terminal to interactively process the power calculation service request with the edge side according to the power calculation indicator of the terminal, the geographic position information of the terminal and the SLA guarantee requirement of the target user, acquiring a KQI index matched with the power calculation service request according to the current network condition of the network to which the edge side belongs and the power calculation indicator of the terminal; and under the condition that the KQI index does not meet the SLA guarantee requirement, generating a calculation service request indicator of the target user and sending the calculation service request indicator to a cloud end, wherein the cloud end is used for cooperatively processing the calculation service request with the terminal according to the user intention, the terminal calculation service request indicator and the calculation service request indicator and providing calculation support of a cloud end large model for the terminal.
  17. 17. The method of claim 16, wherein the method further comprises: And under the condition that the KQI index meets the SLA guarantee requirement, processing the calculation service request interactively with the terminal.
  18. 18. The method according to any of claims 16-17, wherein the KQI indicator comprises at least one indicator of a number of camping, a length of camping, a code rate, a resolution, a response delay, a setup delay, jitter, and a bit error rate.
  19. 19. A method for processing a computing power service request, which is applied to a terminal, the method comprising: Under the condition that a target user initiates a calculation service request, acquiring a terminal calculation indicator and uploading the terminal calculation indicator to a cloud; the cloud end is used for forwarding the terminal computing power indicator to an edge side and indicating the terminal to interact with the edge side to process the computing power service request according to the terminal computing power indicator, the geographic position information of the terminal and the SLA guarantee requirement of the target user; The edge side is used for acquiring a KQI index matched with the power calculation service request according to the current network condition of the network to which the edge side belongs and the terminal power calculation indicator; the cloud end is used for cooperatively processing the computing power service request with the terminal according to the user intention, the terminal computing power indicator and the computing power service request indicator and providing computing power support of a cloud end large model for the terminal.
  20. 20. The method of claim 19, wherein the obtaining the terminal calculation indicator comprises: acquiring communication data, user behavior data, personalized preference data, terminal computing power information, load information and link quality information of the terminal, wherein session scene information of the terminal is acquired; Generating the terminal computing power indicator based on the session scene information, the terminal computing power information, the load information, the link quality information and the computing power service request; the user intent is obtained by: And carrying out local processing on the communication data, the user behavior data and the personalized preference data to obtain the user intention, wherein the local processing comprises personal privacy data protection encryption, semantic query and intention recognition.

Description

Computing power service request processing system, method, device, computer equipment and medium Technical Field The present application relates to the field of wireless communication technology and AI terminal technology, and in particular, to a computing power service request processing system, method, apparatus, computer device, computer readable storage medium, and computer program product. Background The AI terminal supports the end cloud hybrid AI architecture and the end side small model localization deployment, realizes multi-mode content generation, full scene perception, safety and reliability, but has limited model calculation power, is difficult to support trillion-level operation, and has difficult balance of energy efficiency ratio, and the large model operated by the cloud server cluster has infinite calculation power, supports hundred-card cluster training, is subject to network transmission, calculation power tide, and link quality interference, and the like, and causes high delay, thereby making the SLA protocol difficult to reach. The terminal side pursues natural conflicts of low power consumption, high real-time and large-scale delay of the cloud model, so that the processing efficiency of the target computing power service request processing mode is low. Disclosure of Invention In view of the foregoing, it is desirable to provide a computing power service request processing system, method, apparatus, computer device, computer readable storage medium, and computer program product that can improve processing efficiency. In a first aspect, the application provides a computing power service request processing system, which comprises a terminal, an edge side and a cloud end; The terminal is used for acquiring a terminal computing power indicator and uploading the terminal computing power indicator to the cloud end under the condition that a target user initiates a computing power service request; the cloud end is used for forwarding the terminal calculation force indicator to the edge side and indicating the terminal to interactively process the calculation force service request with the edge side according to the terminal calculation force indicator, the geographic position information of the terminal and the SLA guarantee requirement of the target user; The edge side is used for acquiring a KQI index matched with the power calculation service request according to the current network condition of the network to which the edge side belongs and the terminal power calculation indicator; The cloud end is used for cooperatively processing the computing power service request with the terminal according to the user intention, the terminal computing power indicator and the computing power service request indicator and providing computing power support of a cloud end large model for the terminal. In combination with the first aspect, in one embodiment, the cloud end is further configured to determine, under the indication of the terminal power calculation symbol, an anchor UPF corresponding to the terminal according to geographic location information of the terminal and SLA guarantee requirements, so as to instruct the terminal to process the power calculation service request based on the anchor UPF and the edge side interaction. In combination with the first aspect, in one embodiment, the cloud end is further configured to obtain a KQI guarantee required by the target user according to the calculation service request indicator, determine a new anchor UPF from the candidate UPFs according to real-time cell location information of the candidate UPFs in case that the KQI guarantee characterizes that the anchor UPF needs to be replaced, process the calculation service request in conjunction with the terminal according to the new anchor UPF and the user intention, and provide calculation support of the cloud end large model to the terminal, and in case that the KQI guarantee characterizes that the anchor UPF does not need to be replaced, subscribe to the NWDAF to subscribe to the policy suggestion based on the terminal calculation service request indicator and the calculation service request indicator, and send a user session update suggestion to the PCF to enable the PCF to establish association with the new anchor UPF. In combination with the first aspect, in an exemplary embodiment, the cloud end is further configured to report experience data of a new anchor UPF subscription terminal through NWDAF, where the subscription information carries a user identifier, so that the new anchor UPF uploads the experience data of a target user corresponding to the user identifier to the NWDAF, the NWDAF predicts a network support service situation according to the experience data, and generates an initial SLA guarantee policy suggestion and returns to the cloud end based on the predicted network support service situation and a requirement of the terminal; The cloud end is further used for receiving the initial S