WO2020259276A1 - 一种网络优化的方法、装置和无线网络优化控制功能元 - Google Patents

一种网络优化的方法、装置和无线网络优化控制功能元 Download PDF

Info

Publication number
WO2020259276A1
WO2020259276A1 PCT/CN2020/095113 CN2020095113W WO2020259276A1 WO 2020259276 A1 WO2020259276 A1 WO 2020259276A1 CN 2020095113 W CN2020095113 W CN 2020095113W WO 2020259276 A1 WO2020259276 A1 WO 2020259276A1
Authority
WO
WIPO (PCT)
Prior art keywords
rcf
real
wireless
time
optimization
Prior art date
Application number
PCT/CN2020/095113
Other languages
English (en)
French (fr)
Inventor
曾玲玲
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Priority to US17/294,904 priority Critical patent/US20210409974A1/en
Priority to EP20831624.0A priority patent/EP3869849B1/en
Publication of WO2020259276A1 publication Critical patent/WO2020259276A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence

Definitions

  • This article relates to, but is not limited to, a network optimization method, device, RCF (Radio network optimization Control Functional, radio network optimization control functional element) and computer-readable storage media.
  • RCF Radio network optimization Control Functional, radio network optimization control functional element
  • Edge computing technology uses servers at the edge of the wireless network to sink cloud computing resources to the wireless access network, shortening the physical distance between the UE (User Equipment) terminal and the service APP (Application), which can greatly Reducing the delay can save the bandwidth of the backhaul network.
  • Edge computing effectively integrates mobile communication network and Internet technology, and has the characteristics of business localization, close range, and low latency.
  • ETSI European Telecommunications Standards Institute
  • ETSI European Telecommunications Standards Institute
  • -MEC standard defines the edge computing framework as shown in Figure 2.
  • the main functional elements include:
  • MEC host includes MEC platform (MEC platform), MEC APP (MEC application) and virtualization infrastructure.
  • MEC platform is mainly responsible for providing multi-access edge services and collecting necessary operating information for edge applications.
  • the virtualization infrastructure can include a data plane.
  • MEC host level management including MEC platform management, responsible for MEC platform functional meta management, application rules and life cycle management; and virtualization infrastructure management, responsible for allocating, managing and releasing virtualized resources.
  • MEC system level management (MEC system level management): Including edge orchestrator, responsible for selecting MEC host, loading applications, triggering application instance initialization and termination; and operation support system, responsible for operation and maintenance and operation, and requesting terminal equipment and third-party customers Authorization and go to the Edge Orchestrator.
  • edge orchestrator responsible for selecting MEC host, loading applications, triggering application instance initialization and termination
  • operation support system responsible for operation and maintenance and operation, and requesting terminal equipment and third-party customers Authorization and go to the Edge Orchestrator.
  • MEC network including 3GPP (3rd Generation Partnership Project, 3rd Generation Partnership Project) network, local network, external network, etc.
  • ETSI-MEC As the most mainstream reference architecture for edge computing systems, ETSI-MEC is mainly developed around the deployment of MEC in 4G communication networks, and will be integrated with 3GPP 5G communication networks to evolve in the future.
  • the embodiments of the present application provide a network optimization method, device, RCF, and computer-readable storage medium.
  • the embodiment of the present application provides a method for network optimization, including: the wireless network optimization control function element RCF in the edge computing system determines real-time characteristics according to real-time data of the wireless access network; the RCF generates wireless optimization according to the real-time characteristics Auxiliary strategy; the RCF issues the wireless optimization auxiliary strategy.
  • An embodiment of the present application also provides a network optimization device, including: a real-time feature module, used for the RCF in the edge computing system to determine real-time features based on real-time data of the wireless access network; a policy generation module, used for determining the real-time feature based on the real-time feature Generate a wireless optimization auxiliary strategy; a strategy publishing module for publishing the wireless optimization auxiliary strategy.
  • a network optimization device including: a real-time feature module, used for the RCF in the edge computing system to determine real-time features based on real-time data of the wireless access network; a policy generation module, used for determining the real-time feature based on the real-time feature Generate a wireless optimization auxiliary strategy; a strategy publishing module for publishing the wireless optimization auxiliary strategy.
  • An embodiment of the present application also provides an RCF, including: a memory, a processor, and a computer program stored on the memory and capable of running on the processor, and the method is implemented when the processor executes the program.
  • An embodiment of the present application also provides a computer-readable storage medium that stores computer-executable instructions, and the computer-executable instructions are used to execute the method.
  • Figure 1 is the location of MEC in the cloud computing network
  • FIG. 1 is the ETSI-MEC system framework
  • Figure 3 is the overall architecture of the 4G/5G wireless access network
  • Figure 4 is the position of the RCF proposed in the embodiment of the application in the ETSI-MEC system framework
  • Figure 5 is a flowchart of a network optimization method according to an embodiment of the present application.
  • Fig. 6 is a flowchart of a network optimization method according to another embodiment of the present application.
  • FIG. 7 is a schematic diagram of the composition of a system for network optimization based on edge computing according to an embodiment of the present application.
  • FIG. 8 is a flowchart of the RCF obtaining real-time information from a radio access network element according to an embodiment of the present application
  • FIG. 9 is a flow chart of the RCF actively publishing a wireless optimization assistance strategy according to an embodiment of the present application.
  • FIG. 10 is a flowchart of RCF query and issuance of wireless optimization assistance strategies in an embodiment of the present application
  • FIG. 11 is a flowchart in which the RCF of the embodiment of the present application obtains the intelligent offline model from the management system, and reports the evaluation result of the intelligent model application to the management system;
  • FIG. 12 is a flow chart of the RCF of the embodiment of the present application sending a wireless optimization assistance strategy indication to the service APP through the third interface a;
  • FIG. 13 is a flowchart in which the RCF of the embodiment of the present application obtains the intelligent offline model from the intelligent algorithm APP, and reports the evaluation result of the intelligent model application to the intelligent algorithm APP;
  • FIG. 14 is a flowchart of RCF completing feature extraction and intelligent algorithm APP completing online reasoning in an embodiment of this application;
  • FIG. 15 is a flowchart of Application Example 1 of the present application.
  • FIG 16 is a flowchart of Application Example 2 of the present application.
  • FIG. 17 is a schematic diagram of a network optimization device according to an embodiment of the present application.
  • FIG. 18 is a schematic diagram of a network optimization device according to another embodiment of the present application.
  • Intelligence is an important development direction of mobile communication networks.
  • Mobile communication networks should use emerging artificial intelligence and deep learning technologies to form self-adaptation and self-driving forces, reduce network operating costs, and create a new era for network operations.
  • 3GPP's 4/5G wireless access network adopts a flat architecture as shown in Figure 3.
  • MME Mobility Management Entity
  • S-GW Serving Gateway
  • eNB Enhanced Node B
  • S1 interface S1 interface
  • eNB Enhanced Node B
  • X2 interface forming E-UTRAN (Evolved UMTS Terrestrial Radio Access Network)
  • AMF Access and Mobility Management Function
  • UPF User Plane Function
  • gNB New Air Interface Next-Generation Node B
  • 5GC No.
  • 5G flat network including 5th generation core network) and NG-RAN (next generation radio access network).
  • NG-RAN next generation radio access network
  • the embodiments of this application deeply integrate MEC technology and 3GPP's 4/5G wireless access network to form efficient interactions, thereby providing technical support for regionally concentrated and near real-time wireless network intelligent optimization, and at the same time extending the edge computing system to wireless Application areas of network intelligence optimization.
  • the embodiment of the application adds new function elements to the edge computing system, downloads or generates intelligent models based on artificial intelligence or machine learning algorithms, collects real-time data in an area, performs online model reasoning, and publishes wireless optimization assistance strategies.
  • the embodiment of the present application adds RCF in the edge computing system.
  • the RCF can be connected to the MEC management system, the wireless access network elements in the 3GPP network, and the MEC APP.
  • 3GPP's radio access network elements include: eNB (enhanced node B), gNB (new air interface next generation node B) and ng-eNB (enhanced node B supporting NG interface); MEC APP refers to the APP on the server side, usually As a third-party APP, MEC APP can be divided into 2 types, one is business APP, and the other is smart algorithm APP.
  • the network optimization method of the embodiment of the present application includes:
  • Step 101 The RCF in the edge computing system determines real-time features according to real-time data of the radio access network.
  • RCF collects real-time data and extracts real-time features.
  • the real-time data refers to non-historical data, which may include data obtained in time at the current time, and may also include data obtained within a current period of time.
  • the data acquired in the current period of time can also be called near real-time data.
  • the real-time data includes real-time information
  • the RCF obtains real-time information from a radio access network element, and extracts real-time features from the real-time information.
  • the RCF can also extract real-time features from the data stream on the data plane.
  • the method before step 101, the method further includes:
  • Step 100 RCF determines the intelligent model.
  • the RCF may adopt the following three methods to determine the intelligent model:
  • the RCF obtains the intelligent model from the MEC management system.
  • RCF can receive the intelligent offline model that has been trained from the MEC management system.
  • the RCF obtains the intelligent model from the MEC intelligent algorithm APP.
  • RCF receives customized smart offline models from smart algorithm APP.
  • the RCF trains to obtain the intelligent model according to the real-time data obtained by itself.
  • RCF trains and generates a local intelligent model based on the real-time data it collects.
  • the intelligent model is generated based on a certain amount of data using artificial intelligence and machine learning algorithms.
  • the RCF updates a local intelligent model according to the real-time feature.
  • the local intelligent model may be an intelligent offline model.
  • Step 102 The RCF generates a wireless optimization assistance strategy according to the real-time feature.
  • step 102 includes: the RCF is based on an intelligent model, performs online reasoning according to the real-time feature, and generates a wireless optimization assistance strategy.
  • step 102 includes: the RCF reports the real-time feature to the MEC intelligent algorithm APP, and generates a wireless optimization assistance strategy according to the online reasoning result indication provided by the MEC intelligent algorithm APP.
  • Step 103 The RCF publishes the wireless optimization assistance strategy.
  • the RCF sends the radio optimization assistance strategy to a radio access network element, so that the radio access network element realizes radio resource optimization.
  • the RCF sends the wireless optimization auxiliary strategy to the wireless access network element, and the wireless access network element completes the wireless resource optimization according to the policy instruction and in combination with its own wireless resource management algorithm.
  • the RCF sends the radio optimization assistance strategy to the MEC service APP, so that the MEC APP is optimized for UE applications.
  • the RCF sends the wireless optimization assistance strategy to the MEC service APP, and the MEC APP performs optimization for UE applications in accordance with the policy instructions and combined with its own service application layer algorithm.
  • the intelligent offline model downloaded to the local after the step 103, it further includes:
  • Step 104 The RCF evaluates the intelligent model according to the execution of online reasoning.
  • the method further includes:
  • Step 105 The RCF feeds back an evaluation report to the provider of the smart model.
  • the provider of the smart model can restart offline model training accordingly.
  • the embodiment of the application integrates the MEC technology and the wireless access network, and efficiently interacts real-time wireless data and intelligent wireless optimization auxiliary strategies between the two, forming a closed-loop control of wireless network intelligent optimization, and achieving wireless network self-driving optimization.
  • the system for network optimization based on edge computing includes the following modules:
  • FIG 4 shows the position of RCF in the ETSI-MEC edge computing framework. The relationship between RCF and other modules is shown in Figure 7.
  • the interface between RCF24 and 3GPP radio access network element 23 is the first interface. Through the first interface, the RCF obtains real-time information from the wireless access network element, and the RCF sends the wireless optimization auxiliary strategy to the wireless access network element.
  • RCF obtains real-time information from wireless access network elements, including the following steps:
  • Step 301 The RCF sends a real-time information request message to the radio access network element.
  • the message may include, but is not limited to: base station identification, cell identification, collection information type, collection status, etc.
  • Step 302 The RCF receives a real-time information response message from the wireless access network element.
  • the message can include but is not limited to: execution results, etc.
  • Step 303 The RCF receives a real-time information report message from the radio access network element.
  • the message may include but is not limited to: base station identification, cell identification, cell load, neighboring cell interference, neighboring cell load, UE identification, intra-frequency/inter-frequency/inter-system measurement, handover statistics, collection time, event/period type, etc.
  • RCF issues wireless optimization assistance strategies, which can be active release methods or query release methods.
  • active release includes the following steps:
  • Step 401 The RCF sends a wireless optimization assistance strategy indication message to the wireless access network element.
  • the message may include, but is not limited to: base station identification, cell identification, neighboring cell assistance information, UE identification, algorithm type, suggested target cell, suggested Qos (quality level) priority, suggested POLICY (policy) type, etc.
  • the RCF receives a wireless optimization assistance strategy indication confirmation message from the wireless access network element.
  • the message can include but is not limited to: execution results, etc.
  • query publishing can include the following steps:
  • Step 501 The RCF receives a wireless optimization assistance strategy request message sent by the wireless access network element.
  • the message may include but is not limited to: base station identification, UE identification, cell identification and its real-time information, neighboring cell information, etc.
  • Step 502 The RCF sends a radio optimization assistance strategy response message to the radio access network element.
  • the message may include, but is not limited to: base station identification, cell identification, neighboring cell assistance information, UE identification, algorithm type, suggested target cell, suggested Qos (quality level) priority, suggested POLICY (policy) type, etc.
  • the interface between RCF24 and management system 21 is the second interface. Through this second interface, the RCF 24 obtains the trained smart offline model from the management system 21, and the RCF 24 reports the evaluation result of the smart model application to the management system 21. As shown in Figure 11, the following steps may be included:
  • the RCF receives an intelligent model update command message of the management system.
  • the message can include but is not limited to: user name, password, download address, file name, etc.
  • step 602 the RCF starts the transmission process to complete the intelligent model download.
  • This process can include but is not limited to: file transfer process.
  • the content of the file may include, but is not limited to: the type of smart model, the identification of the changed model, model data, etc.
  • Intelligent model types may include, but are not limited to: radio frequency fingerprint library, intelligent positioning fingerprint library, cell interference model, TCP (Transmission Control Protocol, Transmission Control Protocol) layer intelligent optimization model, business intelligent identification model, business intelligent prediction model, etc.
  • TCP Transmission Control Protocol, Transmission Control Protocol
  • Step 603 RCF sends a model update command result message to the management system.
  • the message can include but is not limited to: execution results, etc.
  • the RCF may also send an intelligent model evaluation report message to the management system.
  • the message may include but is not limited to: intelligent model type, model identification, model availability, model execution statistics, etc.
  • the interface between RCF24 and MEC APP22 is the third interface.
  • the third interface a is between the service APP and the third interface b between the intelligent algorithm APP.
  • the RCF sends a wireless optimization assistance strategy instruction to the service APP to trigger the APP to perform service application layer optimization.
  • It can be an active release method or a query release method. As shown in Figure 12, it includes the following steps:
  • Step 701 RCF sends a wireless optimization assistance strategy instruction message to the service APP.
  • the message may include but is not limited to: APP ID, UE ID, recommended video bit rate, etc.
  • the RCF receives a wireless optimization assistance strategy indication confirmation message returned by the service APP.
  • the message may include but is not limited to: APP ID, UE ID, execution result, etc.
  • the RCF receives the wireless optimization assistance strategy request message sent by the service APP.
  • the message may include but is not limited to: APP ID, UE ID, the type of strategy that needs to be recommended, etc.
  • Step 704 The RCF sends a wireless optimization assistance strategy response message to the service APP.
  • the content of the message refer to step 701.
  • the RCF obtains the trained smart offline model from the smart algorithm APP, and the RCF reports the evaluation result of the smart model application to the smart algorithm APP.
  • the RCF can include the following steps:
  • the RCF receives the smart model update command message of the smart algorithm APP.
  • the message can include but is not limited to: user name, password, download address, file name, etc.
  • step 802 RCF starts the transmission process and completes the download of the smart model.
  • This process can include but is not limited to: file transfer process.
  • the content of the file may include, but is not limited to: the type of smart model, the identification of the changed model, model data, etc.
  • Intelligent model types can include but are not limited to: radio frequency fingerprint library, intelligent positioning fingerprint library, cell interference model, TCP layer intelligent optimization model, business intelligent identification model, business intelligent prediction model, etc.
  • Step 803 RCF sends a model update command result message to the smart algorithm APP.
  • the message can include but is not limited to: execution results, etc.
  • the RCF may also send a smart model evaluation report message to the smart algorithm APP.
  • the message may include but is not limited to: intelligent model type, model identification, model availability, model execution statistics, etc.
  • Method II RCF completes feature extraction, APP completes online reasoning.
  • the RCF sends a real-time feature report to the intelligent algorithm APP, and at the same time receives an online reasoning result indication sent by it, and the RCF then generates a wireless optimization auxiliary strategy.
  • it includes the following steps:
  • Step 901 RCF sends a real-time feature report message to the smart algorithm APP.
  • the message may include but is not limited to: APP ID, smart model type, model ID, designated feature statistical value, etc.
  • the RCF receives an online reasoning result indication message of the smart algorithm APP.
  • the message may include but is not limited to: APP identification, intelligent model type, model identification, business prediction results, etc.
  • a global real-time (near real-time) wireless data view can be established in the edge computing area, forming a regionally concentrated near-real-time wireless intelligent optimization control, and giving play to the advantages of centralized resource decision-making and unified scheduling. Ultimately, it can improve wireless resource utilization and enhance user perception experience.
  • the embodiments of this application are compatible with 4/5G wireless access networks.
  • the 4/5G wireless access network element focuses on the 3GPP protocol stack processing.
  • RCF obtains the intelligent offline model from the management system and issues auxiliary strategies to the wireless access network elements. As shown in Figure 15, the main processing steps are as follows:
  • RCF obtains an intelligent offline model from the management system.
  • Step 1001 RCF receives an intelligent model update command from the management system.
  • the message includes but is not limited to: user name, password, download address and file name.
  • Step 1002 RCF starts the transmission process and completes the download of the smart model.
  • This process can include but is not limited to: file transfer process.
  • the content of the file may include, but is not limited to: the type of smart model, the identification of the changed model, and the model data.
  • Step 1003 RCF returns the result of the intelligent model update command to the management system.
  • the information element carried in the message may include but is not limited to: execution result.
  • Steps 1004 to 1006, RCF collects real-time data of wireless access network elements.
  • Step 1004 The RCF sends a real-time information request to the wireless access network element.
  • the message may include, but is not limited to: base station identification, cell identification, collection information type, and collection status.
  • Step 1005 The wireless access network element returns a response.
  • the message can include but is not limited to: execution result.
  • Step 1006 The wireless access network element performs reporting.
  • the message may include but is not limited to: base station identification, cell identification, cell load, neighboring cell interference, neighboring cell load, UE identification, intra-frequency/inter-frequency/inter-system measurement, handover statistics, collection time, event/period type.
  • Steps 1007-1009 RCF extracts real-time features and executes online reasoning.
  • Step 1007 RCF extracts specified real-time features from real-time information and real-time features from the Data plane data stream according to the requirements of the intelligent model.
  • Step 1008 RCF performs online reasoning based on real-time features; at the same time, RCF updates the local intelligent offline model.
  • Step 1009 RCF generates a wireless optimization assistance strategy according to the inference result.
  • the RCF issues a wireless optimization assistance strategy to the wireless access network element.
  • Step 1010 The RCF receives a request for a wireless optimization assistance strategy from the wireless access network element.
  • the message may include but is not limited to: base station identification, UE identification, cell identification and its real-time information, and neighboring cell information.
  • Step 1011 The RCF sends a wireless optimization assistance strategy response to the wireless access network element.
  • the message may include but is not limited to: base station identification, cell identification, neighboring cell assistance information, UE identification, algorithm type, suggested target cell, and suggested QoS priority.
  • Step 1012 RCF reports the intelligent model evaluation to the management system.
  • RCF sends an intelligent model evaluation report to the management system.
  • the message includes but is not limited to: intelligent model type, model identification, model availability, model execution statistics.
  • the management system can determine whether to retrain the intelligent offline model based on this.
  • RCF obtains online reasoning results from smart algorithm APP, and issues auxiliary strategies to wireless access network elements and service APPs. As shown in Figure 16, the main processing steps are as follows:
  • Steps 1101 to 1102 RCF accepts the optimization service request of the business APP.
  • Step 1101 RCF receives an optimization service registration request of the business APP.
  • the message can include but is not limited to: APP ID, UE ID, and service type.
  • Step 1102 RCF sends an optimized service registration response to the service APP.
  • the message can include but is not limited to: APP ID, UE ID, service type, execution result.
  • Steps 1103 to 1104, RCF accepts the algorithm service request of the smart algorithm APP.
  • Step 1103 RCF receives the service registration request message of the smart algorithm APP.
  • the message may include, but is not limited to: APP ID, service type, intelligent model type, model ID, and designated characteristics.
  • Step 1104 RCF sends a service registration response to the smart algorithm APP.
  • the message can include but is not limited to: APP ID, service type, and execution result.
  • Steps 1105 to 1107 RCF collects real-time data of wireless access network elements.
  • Step 1105 The RCF sends a real-time information request to the wireless access network element.
  • the message may include, but is not limited to: base station identification, cell identification, collection information type, and collection status.
  • Step 1106 The wireless access network element returns a response.
  • the message can include but is not limited to: execution result.
  • Step 1107 The wireless access network element performs reporting.
  • the message may include but is not limited to: base station identification, cell identification, cell load, neighboring cell interference, neighboring cell load, UE identification, intra-frequency/inter-frequency/inter-system measurement, handover statistics, collection time, event/period type.
  • Steps 1108 to 1109 RCF reports real-time features to the smart algorithm APP.
  • Step 1108 RCF extracts the specified features from the real-time information, and sends a real-time feature report to the APP.
  • the message may include, but is not limited to: APP identification, intelligent model type, model identification, and designated feature statistical values.
  • Step 1109 RCF receives the online reasoning result indication message sent by the smart algorithm APP.
  • the message may include but is not limited to: APP ID, intelligent model type, model ID, business prediction result.
  • the RCF may issue a wireless optimization auxiliary strategy to the wireless access network element.
  • Step 1110 The RCF sends a wireless optimization assistance strategy instruction to the wireless access network element.
  • the message may include, but is not limited to: base station identity, cell identity, UE identity, algorithm type, and suggested Qos priority.
  • Step 1111 The RCF receives the indication confirmation returned by the wireless access network element.
  • the message can include but is not limited to: execution result.
  • Steps 1112 to 1113 RCF can also issue wireless optimization assistance strategies to the business APP.
  • Step 1112 RCF sends a wireless optimization assistance strategy instruction to the service APP.
  • the message can include but is not limited to: APP ID, UE ID, and recommended video bit rate.
  • Step 1113 RCF receives the indication confirmation returned by the service APP.
  • the message can include but is not limited to: execution result.
  • an embodiment of the present application also provides a network optimization device, including:
  • the real-time feature module 1201 is used for the RCF in the edge computing system to determine the real-time feature according to the real-time data of the wireless access network;
  • the strategy generation module 1202 is configured to generate a wireless optimization auxiliary strategy according to the real-time feature
  • the strategy publishing module 1203 is configured to publish the wireless optimization auxiliary strategy.
  • the real-time data includes real-time information
  • the real-time feature module 1201 is configured to obtain real-time information from a wireless access network element, and extract real-time features from the real-time information.
  • the real-time feature module 1201 is used to extract real-time features from the data stream of the data plane.
  • the device further includes:
  • the intelligent model determination module 1200 is used to determine the intelligent model.
  • the smart model determination module 1200 is configured to determine the smart model in at least one of the following ways:
  • the intelligent model is obtained through training.
  • the smart model determining module 1200 is further configured to update the local smart model according to the real-time feature.
  • the strategy generation module 1202 is used for the RCF based on an intelligent model to perform online reasoning according to the real-time features to generate a wireless optimization assistance strategy.
  • the device further includes:
  • the intelligent model evaluation module 1204 is configured to evaluate the intelligent model according to the execution status of online reasoning.
  • the device further includes:
  • the feedback module 1205 is used to feed back an evaluation report to the provider of the smart model.
  • the strategy generation module 1202 is configured to report the real-time characteristics to the MEC intelligent algorithm APP, and generate the wireless optimization auxiliary strategy according to the online reasoning result indication provided by the MEC intelligent algorithm APP.
  • the policy issuing module 1203 is configured to send the wireless optimization auxiliary policy to a wireless access network element, so that the wireless access network element realizes wireless resource optimization.
  • the policy issuing module 1203 is configured to send the wireless optimization auxiliary policy to the MEC service APP, so that the MEC APP is optimized for user equipment UE applications.
  • An embodiment of the present application also provides an RCF, including: a memory, a processor, and a computer program stored on the memory and capable of running on the processor, and the method for implementing the network optimization when the processor executes the program.
  • An embodiment of the present application also provides a computer-readable storage medium that stores computer-executable instructions, and the computer-executable instructions are used to execute the method for implementing the conference control.
  • the foregoing storage medium may include, but is not limited to: U disk, Read-Only Memory (ROM), Random Access Memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc.
  • U disk Read-Only Memory
  • RAM Random Access Memory
  • RAM Random Access Memory
  • mobile hard disk magnetic disk or optical disk, etc.
  • Such software may be distributed on a computer-readable medium, and the computer-readable medium may include a computer storage medium (or a non-transitory medium) and a communication medium (or a transitory medium).
  • the term computer storage medium includes volatile and non-volatile memory implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Sexual, removable and non-removable media.
  • Computer storage media include but are not limited to RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, magnetic disk storage or other magnetic storage device, or Any other medium used to store desired information and that can be accessed by a computer.
  • communication media usually contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as carrier waves or other transmission mechanisms, and may include any information delivery media .

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

一种网络优化的方法、装置、RCF和计算机可读存储介质，所述方法包括：边缘计算***中的RCF根据无线接入网的实时数据确定实时特征；所述RCF根据所述实时特征生成无线优化辅助策略；所述RCF发布所述无线优化辅助策略。

Description

一种网络优化的方法、装置和无线网络优化控制功能元
相关申请的交叉引用
本申请基于申请号为201910551380.0、申请日为2019年6月24日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此以引入方式并入本申请。
技术领域
本文涉及但不限于一种网络优化的方法、装置、RCF(Radio network optimization Control Functional，无线网络优化控制功能元)和计算机可读存储介质。
背景技术
无线接入技术发展迅猛，移动互联网遍布全球，移动应用越来越丰富，移动通信网络已成为人们生活娱乐不可或缺的必需品。5G(第五代移动通信)时代的三大应用场景，eMBB(Enhance Mobile Broadband，增强型移动宽带)、uRLLC(Ultra Reliable&Low Latency Communication，高可靠低时延通信)和mMTC(Massive Machine Type Communication，大规模机器通信)，将进一步导致移动数据流量和多样性的猛增。
在低时延和高带宽业务不断增长的需求背景之下，MEC(Multi-Access Edge Computing，多接入边缘计算)应运而生。边缘计算技术通过在无线网络边缘配置服务器，将云计算资源下沉至无线接入网，拉近UE(User Equipment，用户设备)终端与业务APP(Application，应用)的物理距离，既能极大降低时延又能节省回传网带宽。边缘计算有效融合了移动通信网和互联网技术，具有业务本地化、近距离、低时延等特点。通过在移动网络中部署边缘计算，可形成分布式云计算架构如图1所示。
ETSI(European Telecommunications Standards Institute，欧洲电信标准协会) -MEC标准定义了边缘计算框架如图2所示。其中，主要功能元包括：
MEC host(MEC主机)：包含MEC platform(MEC平台)、MEC APP(MEC应用)和虚拟化基础设施。MEC平台主要负责提供多接入边缘服务、以及收集边缘应用必要的运行信息。虚拟化基础设施可包含Data plane(数据面)。
MEC host level management(MEC主机级管理)：包括MEC平台管理，负责MEC平台功能元管理、应用的规则和生命周期管理；以及虚拟化基础设施管理，负责分配、管理和释放虚拟化资源。
MEC system level management(MEC***级管理)：包括边缘编排器，负责选择MEC主机、加载应用、触发应用实例初始化和终止；以及运营支撑***，负责运维和运营、对终端设备和第三方客户请求的授权、并转到边缘编排器。
MEC的网络：包括3GPP(3rd Generation Partnership Project，第3代合作伙伴计划)网络、本地网、外部网络等。
ETSI-MEC作为边缘计算***最主流的参考架构，主要围绕4G通信网络部署MEC展开，后续会与3GPP 5G通信网络结合演进。
另一方面，随着日益密集、丰富、高要求的移动应用的增长，移动网络变得越来越复杂。依靠人工进行网络部署操作和网络优化的传统方式将无力应对。
发明内容
以下是对本文详细描述的主题的概述。本概述并非是为了限制权利要求的保护范围。
本申请实施例提供了一种网络优化的方法、装置、RCF和计算机可读存储介质。
本申请实施例提供了一种网络优化的方法，包括：边缘计算***中的无线网络优化控制功能元RCF根据无线接入网的实时数据确定实时特征；所述RCF根据所述实时特征生成无线优化辅助策略；所述RCF发布所述无线优化辅助策略。
本申请实施例还提供一种网络优化的装置，包括：实时特征模块，用于边缘计算***中的RCF根据无线接入网的实时数据确定实时特征；策略生成模块，用于根据所述实时特征生成无线优化辅助策略；策略发布模块，用于发布所述无线优化辅助策略。
本申请实施例还提供一种RCF，包括：存储器、处理器及存储在存储器上 并可在处理器上运行的计算机程序，所述处理器执行所述程序时实现所述方法。
本申请实施例还提供一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令用于执行所述方法。
在阅读并理解了附图和详细描述后，可以明白其他方面。
附图说明
图1是MEC在云计算网络的位置；
图2是ETSI-MEC***框架；
图3是4G/5G无线接入网整体架构；
图4是本申请实施例提出的RCF在ETSI-MEC***框架的位置；
图5是本申请实施例的网络优化的方法的流程图；
图6是本申请另一实施例的网络优化的方法的流程图；
图7是本申请实施例的基于边缘计算的网络优化的***的组成示意图；
图8是本申请实施例的RCF从无线接入网元获取实时信息的流程图；
图9是本申请实施例的RCF主动发布无线优化辅助策略的流程图；
图10是本申请实施例的RCF查询发布无线优化辅助策略的流程图；
图11是本申请实施例的RCF从管理***获取智能离线模型，向管理***上报智能模型应用的评估结果的流程图；
图12是本申请实施例的RCF通过第三接口a向业务APP发送无线优化辅助策略指示的流程图；
图13是本申请实施例的RCF从智能算法APP获取智能离线模型，向智能算法APP上报智能模型应用的评估结果的流程图；
图14是本申请实施例的RCF完成特征提取，智能算法APP完成在线推理的流程图；
图15是本申请应用实例1的流程图；
图16是本申请应用实例2的流程图；
图17是本申请实施例的网络优化的装置的示意图；
图18是本申请另一实施例的网络优化的装置的示意图。
具体实施方式
下文中将结合附图对本申请的实施例进行详细说明。
在附图的流程图示出的步骤可以在诸如一组计算机可执行指令的计算机***中执行。并且，虽然在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤。
智能化是移动通信网络的重要发展方向。移动通信网络要利用新兴的人工智能和深度学习技术，形成自适应和自驱力，降低网络运营成本，为网络运营创造新时代。
为降低用户面时延、避免单点故障，3GPP的4/5G无线接入网络采用扁平化架构如图3所示。其中，MME(移动性管理实体)、S-GW(服务网关)和eNB(增强节点B)之间通过S1接口、eNB之间通过X2接口，形成E-UTRAN(演进的UMTS陆地无线接入网)的扁平网络；AMF(接入和移动管理功能)、UPF(用户面功能)和gNB(新空口下一代节点B)之间通过NG接口，gNB之间直接通过Xn接口，形成包括5GC(第五代核心网)和NG-RAN(下一代无线接入网)在内的5G扁平网络。在扁平化网络中没有分布式计算锚点，无法针对特定区域建立全局实时/近实时数据视图，无法形成区域集中的近实时的智能优化控制，这将成为移动通信网络向智能化全面发展的一大障碍。
本申请实施例将MEC技术和3GPP的4/5G无线接入网进行深度融合、形成高效交互，从而为区域集中的和近实时的无线网络智能优化提供技术支撑，同时将边缘计算***扩展到无线网络智能优化的应用领域。
本申请实施例在边缘计算***中新增功能元，下载或者生成基于人工智能或者机器学习算法的智能模型、收集区域范围的实时数据、执行在线模型推理、发布无线优化辅助策略。
如图4所示，本申请实施例在边缘计算***内增加RCF，RCF可连接MEC管理***、3GPP网络中的无线接入网元、MEC APP。
其中，3GPP的无线接入网元包括：eNB(增强节点B)、gNB(新空口下一代节点B)和ng-eNB(支持NG接口的增强节点B)；MEC APP是指服务器端的APP，通常为第三方APP，MEC APP可分为2类，一类是业务APP，一类是智能算法APP。
如图5所示，本申请实施例的网络优化的方法，包括：
步骤101，边缘计算***中的RCF根据无线接入网的实时数据确定实时特征。
本步骤中，RCF搜集实时数据，提取实时特征。
所述实时数据是指非历史数据，可以包括当前时间及时获取的数据，也可以包括当前一段时间内获取的数据。当前一段时间内获取的数据也可称为近实时数据。
在一实施例中，所述实时数据包括实时信息，所述RCF从无线接入网元获取实时信息，从所述实时信息中提取实时特征。
所述RCF还可以从数据面的数据流提取实时特征。
如图6所示，在一实施例中，所述步骤101之前，还包括：
步骤100，RCF确定智能模型。
其中，所述RCF可以采用如下三种方式确定智能模型：
1、所述RCF从MEC管理***获取所述智能模型。
这种方式中，RCF可以从MEC管理***，接收已完成训练的智能离线模型。
2、所述RCF从MEC智能算法APP获取所述智能模型。
这种方式中，RCF从智能算法APP接收定制化的智能离线模型。
3、所述RCF根据自身获取的实时数据，训练得到所述智能模型。
这种方式中，RCF根据自己搜集的实时数据，训练生成本地智能模型。
其中，智能模型基于一定规模的数据量、采用人工智能和机器学习算法来生成。
在一实施例中，所述RCF根据所述实时特征更新本地的智能模型。
该本地的智能模型可以是智能离线模型。
步骤102，所述RCF根据所述实时特征生成无线优化辅助策略。
在一实施例中，步骤102包括：所述RCF基于智能模型，根据所述实时特征执行在线推理，生成无线优化辅助策略。
在另一实施例中，步骤102包括：所述RCF将所述实时特征上报给MEC智能算法APP，根据所述MEC智能算法APP提供的在线推理结果指示，生成无线优化辅助策略。
步骤103，所述RCF发布所述无线优化辅助策略。
在一实施例中，所述RCF将所述无线优化辅助策略发送至无线接入网元，以使所述无线接入网元实现无线资源优化。
其中，RCF将所述无线优化辅助策略发送至无线接入网元，由无线接入网元根据策略指示，结合自有的无线资源管理算法，完成无线资源优化。
在一实施例中，所述RCF将所述无线优化辅助策略发送至MEC业务APP， 以使所述MEC APP针对UE应用进行优化。
其中，RCF将所述无线优化辅助策略发送至MEC业务APP，由MEC APP根据策略指示，结合自有的业务应用层算法，针对UE应用进行优化。
如图6所示，在一实施例中，对于下载到本地的智能离线模型，所述步骤103之后，还包括：
步骤104，所述RCF根据在线推理的执行情况，评估所述智能模型。
在一实施例中，所述RCF根据在线推理的执行情况，评估所述智能模型之后，所述方法还包括：
步骤105，所述RCF向所述智能模型的提供者反馈评估报告。
所述智能模型的提供者可据此重启离线模型训练。
本申请实施例融合了MEC技术和无线接入网，在两者之间高效交互实时无线数据和智能无线优化辅助策略，形成无线网络智能优化的闭环控制，达到无线网络自驱优化。
如图7所示，本申请实施例基于边缘计算的网络优化的***包括以下模块：
MEC中的管理***21、MEC APP22、MEC网络中的3GPP无线接入网元23。还包括新增模块：MEC中的RCF24。
RCF在ETSI-MEC边缘计算框架的位置如图4所示。RCF和其他模块的关系如图7所示。
一、RCF24和3GPP无线接入网元23之间的接口为第一接口。通过该第一接口，RCF从无线接入网元获取实时信息，RCF向无线接入网元发送无线优化辅助策略。
如图8所示，RCF从无线接入网元获取实时信息，包括如下步骤：
步骤301，RCF向无线接入网元发送实时信息请求消息。消息中可包含但不限于：基站标识、小区标识、采集信息类型、采集状态等。
步骤302，RCF收到无线接入网元的实时信息响应消息。消息中可包含但不限于：执行结果等。
步骤303，RCF收到无线接入网元的实时信息报告消息。消息中可包含但不限于：基站标识、小区标识、小区负荷、邻区干扰、邻区负荷、UE标识、频内/频间/***间测量、切换统计、采集时间、事件/周期类型等。
RCF发布无线优化辅助策略，可为主动发布方式、或者查询发布方式。
如图9所示，主动发布包括如下步骤：
步骤401，RCF向无线接入网元发送无线优化辅助策略指示消息。消息中可包含但不限于：基站标识、小区标识、邻区辅助信息、UE标识、算法类型、建议的目标小区、建议的Qos(质量等级)优先级、建议的POLICY(策略)类型等。
步骤402，RCF收到无线接入网元的无线优化辅助策略指示确认消息。消息中可包含但不限于：执行结果等。
如图10所示，查询发布可包括如下步骤：
步骤501，RCF收到无线接入网元发送的无线优化辅助策略请求消息。消息中可包含但不限于：基站标识、UE标识、小区标识及其实时信息、邻区信息等。
步骤502，RCF向无线接入网元发送无线优化辅助策略响应消息。消息内容参考步骤401中所述。消息中可包含但不限于：基站标识、小区标识、邻区辅助信息、UE标识、算法类型、建议的目标小区、建议的Qos(质量等级)优先级、建议的POLICY(策略)类型等。
二、RCF24和管理***21之间的接口为第二接口。通过该第二接口，RCF24从管理***21获取训练好的智能离线模型，RCF24向管理***21上报智能模型应用的评估结果。如图11所示，可包括如下步骤：
步骤601，RCF收到管理***的智能模型更新命令消息。消息中可包含但不限于：用户名、密码、下载地址和文件名等。
步骤602，RCF启动传输过程，完成智能模型下载。该过程可包含但不限于：文件传输过程。
文件内容可包含但不限于：智能模型类型、发生变更的模型的标识、模型数据等。
智能模型类型可包含但不限于：无线射频指纹库、智能定位指纹库、小区干扰模型、TCP(Transmission Control Protocol，传输控制协议)层智能优化模型、业务智能识别模型、业务智能预测模型等。
步骤603，RCF向管理***发送模型更新命令结果消息。消息中可包含但不限于：执行结果等。
步骤604，RCF还可以向管理***发送智能模型评估报告消息。消息中可包含但不限于：智能模型类型、模型标识、模型可用程度、模型执行统计等。
三、RCF24和MEC APP22之间的接口为第三接口。其中，和业务APP之间为第三接口a，和智能算法APP之间为第三接口b。
通过第三接口a，RCF向业务APP发送无线优化辅助策略指示，触发APP执行业务应用层优化。可为主动发布方式、或者查询发布方式。如图12所示，包括如下步骤：
主动发布：
步骤701，RCF向业务APP发送无线优化辅助策略指示消息。消息中可包含但不限于：APP标识、UE标识、建议的视频码率等。
步骤702，RCF收到业务APP返回的无线优化辅助策略指示确认消息。消息中可包含但不限于：APP标识、UE标识、执行结果等。
或者查询发布：
步骤703，RCF收到业务APP发送的无线优化辅助策略请求消息。消息中可包含但不限于：APP标识、UE标识、需要建议的策略类型等。
步骤704，RCF向业务APP发送无线优化辅助策略响应消息。消息内容参考步骤701中所述。
第三接口b有2种方式可选，方式I或者方式II。
方式I、RCF完成特征提取和在线推理。
此时通过第三接口b，RCF从智能算法APP获取训练好的智能离线模型，RCF向智能算法APP上报智能模型应用的评估结果。如图13所示，与第二接口类似，可包括如下步骤：
步骤801，RCF收到智能算法APP的智能模型更新命令消息。消息中可包含但不限于：用户名、密码、下载地址和文件名等。
步骤802，RCF启动传输过程，完成智能模型下载。该过程可包含但不限于：文件传输过程。
文件内容可包含但不限于：智能模型类型、发生变更的模型的标识、模型数据等。
智能模型类型可包含但不限于：无线射频指纹库、智能定位指纹库、小区干扰模型、TCP层智能优化模型、业务智能识别模型、业务智能预测模型等。
步骤803，RCF向智能算法APP发送模型更新命令结果消息。消息中可包 含但不限于：执行结果等。
步骤804，RCF还可以向智能算法APP发送智能模型评估报告消息。消息中可包含但不限于：智能模型类型、模型标识、模型可用程度、模型执行统计等。
方式II、RCF完成特征提取、APP完成在线推理。此时通过第三接口b，RCF向智能算法APP发送实时特征报告、同时接收其发送的在线推理结果指示，RCF进而生成无线优化辅助策略。如图14所示，包括如下步骤：
步骤901，RCF向智能算法APP发送实时特征报告消息。消息中可包含但不限于：APP标识、智能模型类型、模型标识、指定特征统计值等。
步骤902，RCF收到智能算法APP的在线推理结果指示消息。消息中可包含但不限于：APP标识、智能模型类型、模型标识、业务预测结果等。
综上所述，通过本申请实施例，可在边缘计算区域内建立全局实时(近实时)无线数据视图，形成区域集中的近实时的无线智能优化控制，发挥资源集中决策和统一调度的优势，最终能提高无线资源利用率和提升用户感知体验。而且，本申请实施例可兼容4/5G无线接入网。4/5G无线接入网元聚焦于3GPP协议栈处理，其无需频繁升级版本、无需引入智能模型获取/解析和实时特征提取的性能开销、无需面对直接向第三方MEC APP开放的安全风险，就可以完成无线优化算法增强，推动整个无线网络的智能优化功能向前演进。
下面以一些应用实例进行说明。
应用实例1：
RCF从管理***获取智能离线模型、向无线接入网元发布辅助策略。如图15所示，主要处理步骤如下：
步骤1001～1003，RCF从管理***获取智能离线模型。
步骤1001，RCF收到管理***的智能模型更新命令。消息中包含但不限于：用户名、密码、下载地址和文件名。
步骤1002，RCF启动传输过程，完成智能模型下载。该过程可包含但不限于：文件传输过程。文件内容可包含但不限于：智能模型类型、发生变更的模型的标识、模型数据。
步骤1003，RCF向管理***返回智能模型更新命令结果。消息中携带的信元可包含但不限于：执行结果。
步骤1004～1006，RCF采集无线接入网元的实时数据。
步骤1004，RCF向无线接入网元发送实时信息请求。消息中可包含但不限于：基站标识、小区标识、采集信息类型、采集状态。
步骤1005，无线接入网元返回响应。消息中可包含但不限于：执行结果。
步骤1006，无线接入网元执行上报。消息中可包含但不限于：基站标识、小区标识、小区负荷、邻区干扰、邻区负荷、UE标识、频内/频间/***间测量、切换统计、采集时间、事件/周期类型。
步骤1007～1009，RCF提取实时特征、执行在线推理。
步骤1007，RCF根据智能模型要求，从实时信息中提取指定的实时特征，从Data plane数据流提取实时特征。
步骤1008，RCF根据实时特征，执行在线推理；同时RCF更新本地智能离线模型。
步骤1009，RCF根据推理结果生成无线优化辅助策略。
步骤1010～1011，RCF向无线接入网元发布无线优化辅助策略。
步骤1010，RCF收到无线接入网元的无线优化辅助策略请求。消息中可包含但不限于：基站标识、UE标识、小区标识及其实时信息、邻区信息。
步骤1011，RCF向无线接入网元发送无线优化辅助策略响应。消息中可包含但不限于：基站标识、小区标识、邻区辅助信息、UE标识、算法类型、建议的目标小区、建议的Qos优先级。
步骤1012，RCF向管理***上报智能模型评估。
其中，RCF向管理***发送智能模型评估报告，消息中包含但不限于：智能模型类型、模型标识、模型可用程度、模型执行统计。管理***可据此判断是否重新训练智能离线模型。
应用实例2：
RCF从智能算法APP获取在线推理结果、向无线接入网元和业务APP发布辅助策略。如图16所示，主要处理步骤如下：
步骤1101～1102，RCF受理业务APP的优化服务请求。
步骤1101，RCF收到业务APP的优化服务注册请求。消息中可包含但不限 于：APP标识、UE标识、服务类型。
步骤1102，RCF向业务APP发送优化服务注册响应。消息中可包含但不限于：APP标识、UE标识、服务类型、执行结果。
步骤1103～1104，RCF受理智能算法APP的算法服务请求。
步骤1103，RCF收到智能算法APP的服务注册请求消息。消息中可包含但不限于：APP标识、服务类型、智能模型类型、模型标识、指定特征。
步骤1104，RCF向智能算法APP发送服务注册响应。消息中可包含但不限于：APP标识、服务类型、执行结果。
步骤1105～1107，RCF采集无线接入网元的实时数据。
步骤1105，RCF向无线接入网元发送实时信息请求。消息中可包含但不限于：基站标识、小区标识、采集信息类型、采集状态。
步骤1106，无线接入网元返回响应。消息中可包含但不限于：执行结果。
步骤1107，无线接入网元执行上报。消息中可包含但不限于：基站标识、小区标识、小区负荷、邻区干扰、邻区负荷、UE标识、频内/频间/***间测量、切换统计、采集时间、事件/周期类型。
步骤1108～1109，RCF向智能算法APP上报实时特征。
步骤1108，RCF从实时信息中提取指定的特征，向APP发送实时特征报告。消息中可包含但不限于：APP标识、智能模型类型、模型标识、指定特征统计值。
步骤1109，RCF收到智能算法APP发送的在线推理结果指示消息。消息可包含但不限于：APP标识、智能模型类型、模型标识、业务预测结果。
步骤1110～1111，RCF可向无线接入网元发布无线优化辅助策略。
步骤1110，RCF向无线接入网元发送无线优化辅助策略指示。消息中可包含但不限于：基站标识、小区标识、UE标识、算法类型、建议的Qos优先级。
步骤1111，RCF收到无线接入网元返回的指示确认。消息中可包含但不限于：执行结果。
步骤1112～1113，RCF也可向业务APP发布无线优化辅助策略。
步骤1112，RCF向业务APP发送无线优化辅助策略指示。消息中可包含但不限于：APP标识、UE标识、建议的视频码率。
步骤1113，RCF收到业务APP返回的指示确认。消息中可包含但不限于：执行结果。
如图17所示，本申请实施例还提供一种网络优化的装置，包括：
实时特征模块1201，用于边缘计算***中的RCF根据无线接入网的实时数据确定实时特征；
策略生成模块1202，用于根据所述实时特征生成无线优化辅助策略；
策略发布模块1203，用于发布所述无线优化辅助策略。
在一实施例中，所述实时数据包括实时信息，所述实时特征模块1201，用于从无线接入网元获取实时信息，从所述实时信息中提取实时特征。
在一实施例中，所述实时特征模块1201，用于从数据面的数据流提取实时特征。
如图18所示，在一实施例中，所述装置还包括：
智能模型确定模块1200，用于确定智能模型。
在一实施例中，所述智能模型确定模块1200，用于采用如下方式中的至少之一确定所述智能模型：
从MEC管理***获取所述智能模型；
从MEC智能算法应用APP获取所述智能模型；
根据自身获取的实时数据，训练得到所述智能模型。
在一实施例中，所述智能模型确定模块1200，还用于根据所述实时特征更新本地的智能模型。
在一实施例中，所述策略生成模块1202，用于所述RCF基于智能模型，根据所述实时特征执行在线推理，生成无线优化辅助策略。
如图18所示，在一实施例中，所述装置还包括：
智能模型评估模块1204，用于根据在线推理的执行情况，评估所述智能模型。
在一实施例中，所述装置还包括：
反馈模块1205，用于向所述智能模型的提供者反馈评估报告。
在一实施例中，策略生成模块1202，用于将所述实时特征上报给MEC智能算法APP，根据所述MEC智能算法APP提供的在线推理结果指示，生成无线优化辅助策略。
在一实施例中，策略发布模块1203，用于将所述无线优化辅助策略发送至 无线接入网元，以使所述无线接入网元实现无线资源优化。
在一实施例中，策略发布模块1203，用于将所述无线优化辅助策略发送至MEC业务APP，以使所述MEC APP针对用户设备UE应用进行优化。
本申请实施例还提供一种RCF，包括：存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述程序时实现所述网络优化的方法。
本申请实施例还提供一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令用于执行所述会议控制的实现方法。
在本实施例中，上述存储介质可以包括但不限于：U盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、移动硬盘、磁碟或者光盘等各种可以存储程序代码的介质。
本领域普通技术人员可以理解，上文中所公开方法中的全部或某些步骤、***、装置中的功能模块/单元可以被实施为软件、固件、硬件及其适当的组合。在硬件实施方式中，在以上描述中提及的功能模块/单元之间的划分不一定对应于物理组件的划分；例如，一个物理组件可以具有多个功能，或者一个功能或步骤可以由若干物理组件合作执行。某些组件或所有组件可以被实施为由处理器，如数字信号处理器或微处理器执行的软件，或者被实施为硬件，或者被实施为集成电路，如专用集成电路。这样的软件可以分布在计算机可读介质上，计算机可读介质可以包括计算机存储介质(或非暂时性介质)和通信介质(或暂时性介质)。如本领域普通技术人员公知的，术语计算机存储介质包括在用于存储信息(诸如计算机可读指令、数据结构、程序模块或其他数据)的任何方法或技术中实施的易失性和非易失性、可移除和不可移除介质。计算机存储介质包括但不限于RAM、ROM、EEPROM、闪存或其他存储器技术、CD-ROM、数字多功能盘(DVD)或其他光盘存储、磁盒、磁带、磁盘存储或其他磁存储装置、或者可以用于存储期望的信息并且可以被计算机访问的任何其他的介质。此外，本领域普通技术人员公知的是，通信介质通常包含计算机可读指令、数据结构、程序模块或者诸如载波或其他传输机制之类的调制数据信号中的其他数据，并且可包括任何信息递送介质。

Claims (15)

  1. 一种网络优化的方法，包括：
    边缘计算***中的无线网络优化控制功能元RCF根据无线接入网的实时数据确定实时特征；
    所述RCF根据所述实时特征生成无线优化辅助策略；
    所述RCF发布所述无线优化辅助策略。
  2. 如权利要求1所述的方法，其中，所述实时数据包括实时信息，所述RCF根据无线接入网的实时数据确定实时特征，包括：
    所述RCF从无线接入网元获取实时信息，从所述实时信息中提取实时特征。
  3. 如权利要求1所述的方法，其中，所述RCF根据无线接入网的实时数据确定实时特征，包括：
    所述RCF从数据面的数据流提取实时特征。
  4. 如权利要求1所述的方法，其中，所述边缘计算***中的RCF根据无线接入网的实时数据确定实时特征之前，所述方法还包括：
    所述RCF确定智能模型。
  5. 如权利要求4所述的方法，其中，所述RCF采用如下方式中的至少之一确定所述智能模型：
    所述RCF从多接入边缘计算MEC管理***获取所述智能模型；
    所述RCF从MEC智能算法应用APP获取所述智能模型；
    所述RCF根据自身获取的实时数据，训练得到所述智能模型。
  6. 如权利要求1所述的方法，其中，所述RCF确定实时特征之后，所述方法还包括：
    所述RCF根据所述实时特征更新本地的智能模型。
  7. 如权利要求1所述的方法，其中，所述RCF根据所述实时特征生成无线优化辅助策略，包括：
    所述RCF基于智能模型，根据所述实时特征执行在线推理，生成无线优化辅助策略。
  8. 如权利要求7所述的方法，其中，所述RCF发布所述无线优化辅助策略之后，所述方法还包括：
    所述RCF根据在线推理的执行情况，评估所述智能模型。
  9. 如权利要求8所述的方法，其中，所述RCF根据在线推理的执行情况，评估所述智能模型之后，所述方法还包括：
    所述RCF向所述智能模型的提供者反馈评估报告。
  10. 如权利要求1所述的方法，其中，所述RCF根据所述实时特征生成无线优化辅助策略，包括：
    所述RCF将所述实时特征上报给MEC智能算法APP，根据所述MEC智能算法APP提供的在线推理结果指示，生成无线优化辅助策略。
  11. 如权利要求1所述的方法，其中，所述RCF发布所述无线优化辅助策略，包括：
    所述RCF将所述无线优化辅助策略发送至无线接入网元，以使所述无线接入网元实现无线资源优化。
  12. 如权利要求1所述的方法，其中，所述RCF发布所述无线优化辅助策略，包括：
    所述RCF将所述无线优化辅助策略发送至MEC业务APP，以使所述MEC APP针对用户设备UE应用进行优化。
  13. 一种网络优化的装置，包括：
    实时特征模块，用于边缘计算***中的RCF根据无线接入网的实时数据确定实时特征；
    策略生成模块，用于根据所述实时特征生成无线优化辅助策略；
    策略发布模块，用于发布所述无线优化辅助策略。
  14. 一种RCF，包括：存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，其中，所述处理器执行所述程序时实现如权利要求1～12 中任意一项所述的方法。
  15. 一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令用于执行权利要求1～12中任意一项所述的方法。
PCT/CN2020/095113 2019-06-24 2020-06-09 一种网络优化的方法、装置和无线网络优化控制功能元 WO2020259276A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US17/294,904 US20210409974A1 (en) 2019-06-24 2020-06-09 Method and apparatus for network optimization and radio network optimization control functional unit (rcf)
EP20831624.0A EP3869849B1 (en) 2019-06-24 2020-06-09 Method, apparatus and radio network optimization control function unit (rcf) for network optimization

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910551380.0 2019-06-24
CN201910551380.0A CN112135305B (zh) 2019-06-24 2019-06-24 一种网络优化的方法、装置和无线网络优化控制功能元

Publications (1)

Publication Number Publication Date
WO2020259276A1 true WO2020259276A1 (zh) 2020-12-30

Family

ID=73849815

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/095113 WO2020259276A1 (zh) 2019-06-24 2020-06-09 一种网络优化的方法、装置和无线网络优化控制功能元

Country Status (4)

Country Link
US (1) US20210409974A1 (zh)
EP (1) EP3869849B1 (zh)
CN (1) CN112135305B (zh)
WO (1) WO2020259276A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US12041469B2 (en) * 2019-09-12 2024-07-16 Telefonaktiebolaget Lm Ericsson (Publ) Data sharing between a Non-RT-RIC and a NearRT-RIC for radio resource management
WO2022164732A1 (en) * 2021-01-29 2022-08-04 Assia Spe, Llc System and method for network and computation performance probing for edge computing

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105898762A (zh) * 2015-01-26 2016-08-24 华为技术有限公司 基站优化及部署方法和装置
CN107846270A (zh) * 2016-09-20 2018-03-27 ***通信有限公司研究院 传输策略配置方法及装置、信息传输方法及装置
CN109561444A (zh) * 2017-09-26 2019-04-02 ***通信有限公司研究院 一种无线数据处理方法及***

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3136790B1 (en) * 2014-05-30 2021-05-05 Huawei Technologies Co., Ltd. Radio network control method and radio network controller
CN107018534A (zh) * 2016-01-28 2017-08-04 中兴通讯股份有限公司 一种实现移动边缘计算服务的方法、装置及***
WO2017220158A1 (en) * 2016-06-24 2017-12-28 Nokia Solutions And Networks Oy Policy control of mobile edge applications
WO2018224984A1 (en) * 2017-06-06 2018-12-13 Telefonaktiebolaget Lm Ericsson (Publ) A method for establishing a connection between a neutral host network and one or more virtual radio access networks
US11244242B2 (en) * 2018-09-07 2022-02-08 Intel Corporation Technologies for distributing gradient descent computation in a heterogeneous multi-access edge computing (MEC) networks
CN109688597B (zh) * 2018-12-18 2020-09-01 北京邮电大学 一种基于人工智能的雾无线接入网络组网方法及装置
US11373099B2 (en) * 2018-12-28 2022-06-28 Intel Corporation Artificial intelligence inference architecture with hardware acceleration
WO2020185130A1 (en) * 2019-03-08 2020-09-17 Telefonaktiebolaget Lm Ericsson (Publ) Dynamic access network selection based on application orchestration information in an edge cloud system
US20220159525A1 (en) * 2019-05-24 2022-05-19 Apple Inc. 5g new radio load balancing and mobility robustness

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105898762A (zh) * 2015-01-26 2016-08-24 华为技术有限公司 基站优化及部署方法和装置
CN107846270A (zh) * 2016-09-20 2018-03-27 ***通信有限公司研究院 传输策略配置方法及装置、信息传输方法及装置
CN109561444A (zh) * 2017-09-26 2019-04-02 ***通信有限公司研究院 一种无线数据处理方法及***

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
HUAWEI ET AL.: "Discussion on supporting MEC with CAPIF", 3GPP TSG-SA WG6 MEETING #25 S6-181078, 27 July 2018 (2018-07-27), XP051471196, DOI: 20200812091158A *
See also references of EP3869849A4 *

Also Published As

Publication number Publication date
US20210409974A1 (en) 2021-12-30
CN112135305A (zh) 2020-12-25
EP3869849A1 (en) 2021-08-25
EP3869849B1 (en) 2024-07-31
EP3869849A4 (en) 2021-12-29
CN112135305B (zh) 2024-04-16

Similar Documents

Publication Publication Date Title
CN108174421B (zh) 一种5g网络中基于mec辅助的数据分流方法
Niknam et al. Intelligent O-RAN for beyond 5G and 6G wireless networks
US11297685B2 (en) System and method for session relocation at edge networks
CN108322937B (zh) 无线接入网中用于网络切片的资源分配方法和编排器
CN113573331B (zh) 一种通信方法、装置及***
CN110831261B (zh) 用于组合的rrc非活动恢复、rrc rna＆nas注册过程的装置
US20210204148A1 (en) Real-time intelligent ran controller to support self-driving open ran
JP6730511B2 (ja) ネットワークポリシー更新のトリガー方法、管理機能エンティティおよびコアネットワークデバイス
WO2020259276A1 (zh) 一种网络优化的方法、装置和无线网络优化控制功能元
CN111200810B (zh) 终端的能力信息的获取方法、装置及***
WO2019129300A1 (zh) 缓存决策方法及装置
CN112236989A (zh) 用于无线网络的移动边缘计算应用管理
US11929938B2 (en) Evaluating overall network resource congestion before scaling a network slice
CN113766576B (zh) 服务质量管理方法、电子设备以及存储介质
CN111935205B (zh) 雾计算网络中基于交替方向乘子法的分布式资源分配方法
US20210297832A1 (en) Facilitating enablement of intelligent service aware access utilizing multiaccess edge computing in advanced networks
US20220038534A1 (en) Methods and system for training and reinforcing computer vision models using distributed computing
KR102071024B1 (ko) 위성 기반 인터넷 액세스 및 전송을 위한 수락 제어 시스템
CN109348486A (zh) 一种异构无线网络资源分配方法
JP6247767B2 (ja) サービス先取り方法、装置および基地局
KR20210046179A (ko) 네트워크 기능 제어방법 및 장치
US20240224120A1 (en) Selective centralized unit-user plane extended buffering
WO2024038555A1 (ja) システム、装置、方法、及び非一時的なコンピュータ可読媒体
WO2021164428A1 (zh) 通信方法、装置及***
Wang Fast and Energy-Efficient Mobility Management in Mobile Edge Computing Networks

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20831624

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2020831624

Country of ref document: EP

Effective date: 20210520

NENP Non-entry into the national phase

Ref country code: DE