Network slice management method, apparatus, device, and storage medium

Abstract

The application discloses a network slice management method, device, equipment and storage medium. The method comprises the following steps: acquiring historical service demand data and a first decision result corresponding to multiple network slice openings, and forming a first data set; acquiring historical network slice performance monitoring data and a second decision result corresponding to multiple network slice operation and maintenance, and forming a second data set; and training a correlation model based on the first data set and the second data set, wherein the correlation model represents the correlation between network quality data and network parameters of the network slice, and then the automatic opening and / or automatic operation and maintenance after opening of the network slice can be realized based on the correlation model, so that the customized demand of users can be met, the efficiency of the opening and / or operation and maintenance of the network slice is improved, the strategy optimization of network resources and network configuration is realized, and the waste of network resources is reduced.

Description

Technical Field

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

[0002] The fifth-generation (5G) mobile communication system introduced the concept of network slices (NS) to address the different network performance requirements of various communication services. A network slice is a logical network that can provide specific network functions and meet specific service requirements. It includes a collection of network function modules and resources required to complete a specific function or service.

[0003] A network slice can be a complete end-to-end network including terminals, access network (AN), transport network (TN), core network (CN), and application servers, providing comprehensive communication services. Alternatively, a network slice can be a combination of terminals, AN, TN, CN, and application servers. A network slice can be divided into several network slice subnet instances (NSSIs). A network slice subnet does not necessarily need to provide complete end-to-end network services. A network slice subnet can be a collection of network functions from the same equipment vendor within the network slice, or a collection of network functions divided by domain, such as a CN network slice subnet, an AN network slice subnet, or a collection composed of deployment locations or other methods. A network slice subnet may be shared by multiple network slices.

[0004] 5G network slicing technology can combine different network resources and configurations to provide vertical industry customers with end-to-end customized dedicated logical networks that offer higher bandwidth, lower latency, higher reliability, and larger-scale connectivity. Service Level Agreements (SLAs) are signed with customers to continuously guarantee business needs. However, this end-to-end customization and SLA guarantee also place higher demands on network design and operation.

[0005] In the current network slicing management process, due to the lack of reasonable and clear design basis and optimization rules, the planning of network parameters during the deployment process and the continuous optimization and protection after the network slice is activated mainly rely on the experience of network designers and network operation and maintenance personnel, which is labor-intensive and inefficient.

[0006] In related technologies, to achieve intelligent management of network slices, network activation and changes are often carried out by pre-designing and deploying templates. These templates contain network resource information, configuration information, and network capabilities (latency, speed, etc.) that have been tested and verified. However, these templates generally only correspond to standard products, providing customers with several network slices of different typical network capability levels, which cannot meet all customized needs or the need to minimize the waste of network resources.

[0007] For example, during network operation optimization, the goal is to predict service-level performance metrics by monitoring data, compare the predictions with the network performance requirements proposed by the customer, and map the performance degradation to different strategies according to pre-defined rules to achieve automation. However, summarizing the rules of performance mapping strategies into expert experience is difficult because customers focus on end-to-end performance metrics such as latency and average uplink / downlink rates, which may involve combinations of multiple strategies across AN, TN, and CN subdomains. This is more complex than optimizing single-domain performance metrics in traditional networks, and different strategies have different degrees of impact on the same performance metric. Overusing resources and configuring optimization strategies can lead to resource waste.

[0008] In summary, how to automatically activate network instances that meet users' customized business needs by combining network resources and network configurations, and how to automatically combine various network policies during subsequent business operations to continuously ensure the quality of the communication services contracted by customers while avoiding resource waste, remains a difficult problem to solve. Summary of the Invention

[0009] In view of this, embodiments of this application provide a network slice management method, apparatus, device, and storage medium, which aim to improve the efficiency of network slice activation and / or operation and maintenance, and effectively avoid network resource waste.

[0010] The technical solution of this application embodiment is implemented as follows:

[0011] In a first aspect, embodiments of this application provide a network slice management method, including:

[0012] Obtain historical service demand data and initial decision results corresponding to multiple network slice activations to form the first dataset;

[0013] Obtain historical network slice performance monitoring data and second decision results corresponding to multiple network slice operation and maintenance, and form a second dataset;

[0014] A correlation model is trained based on the first dataset and the second dataset. The correlation model characterizes the correlation between network quality data and network parameters of network slices.

[0015] In the above scheme, training and generating a correlation model based on the first dataset and the second dataset includes:

[0016] The first and second datasets are transformed and merged according to the set transformation rules to obtain the merged third dataset;

[0017] The correlation model is obtained by training the third dataset based on a pre-established network model.

[0018] The historical service demand data and the historical network slice performance monitoring data are combined into the network quality data of the third dataset, and the first decision result and the second decision result are combined into the network parameters of the third dataset.

[0019] In the above scheme, the pre-established network model is a back propagation (BP) neural network model, and the process of training the third dataset based on the pre-established network model to obtain the correlation model includes:

[0020] The weight values ​​of the BP neural network model are trained bidirectionally based on the network quality data and network parameters of the third dataset to obtain the correlation model.

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

[0022] The correlation model is sent to the network slice management function device.

[0023] Secondly, embodiments of this application provide a network slice management method, the method comprising:

[0024] Receive management request data for managing network slices;

[0025] The management request data is converted into network quality data of the network slice based on the set conversion rules;

[0026] The network quality data is input into the correlation model generated by the method described in the first aspect to obtain the corresponding network parameters;

[0027] Network slices are managed based on the aforementioned network parameters.

[0028] In the above scheme, the management request data is the service requirement data requesting the activation of network slices; the management of network slices based on the network parameters includes:

[0029] The network slice is activated based on the network parameters.

[0030] In the above scheme, the step of enabling network slicing based on the network parameters includes:

[0031] Based on the network parameters, the required information of the network slice subnet is determined, and a network slice instance identifier is generated;

[0032] The network slice subnet requirement information and the network slice instance identifier are sent to the network slice subnet management function device.

[0033] In the above scheme, the management request data is network slice performance monitoring data during the operation of the activated network slice; the management of network slices based on the network parameters includes:

[0034] The network slice is operated and maintained based on the network parameters.

[0035] In the above scheme, the operation and maintenance of the network slice based on the network parameters includes:

[0036] Adjust the resources and configuration of the network slice based on the network parameters;

[0037] If it is determined that the adjusted network handover performance monitoring data of the network slice does not meet the target parameters of the SLA, then the network parameters are adjusted based on the connection weights of the correlation model until the network handover performance monitoring data of the network slice meets the target parameters of the SLA.

[0038] Thirdly, embodiments of this application provide a network slice management device, including:

[0039] The first acquisition module is used to acquire historical service demand data and first decision results corresponding to multiple network slice activations, forming the first dataset.

[0040] The second acquisition module is used to acquire historical network slice performance monitoring data and second decision results corresponding to multiple network slice operation and maintenance, forming a second dataset.

[0041] The correlation model generation module is used to train and generate a correlation model based on the first dataset and the second dataset. The correlation model represents the correlation between the network quality data and network parameters of the network slice.

[0042] Fourthly, embodiments of this application provide a network slice management device, including:

[0043] The receiving module is used to receive management request data for managing network slices;

[0044] The conversion module is used to convert the management request data into network quality data of the network slice based on the set conversion rules;

[0045] The determination module is used to input the network quality data into the correlation model generated by the method described in the first aspect to obtain the corresponding network parameters;

[0046] The management module is used to manage network slices based on the network parameters.

[0047] Fifthly, embodiments of this application provide a network slicing data processing device, including: a processor and a memory for storing a computer program capable of running on the processor, wherein the processor, when running the computer program, executes the steps of the method described in the first aspect of embodiments of this application.

[0048] In a sixth aspect, embodiments of this application provide a network slicing management function device, including: a processor and a memory for storing a computer program capable of running on the processor, wherein the processor, when running the computer program, executes the steps of the method described in the second aspect of embodiments of this application.

[0049] In a seventh aspect, embodiments of this application provide a storage medium storing a computer program, which, when executed by a processor, implements the steps of the method described in the first or second aspect of embodiments of this application.

[0050] The technical solution provided in this application involves acquiring historical service demand data and first decision results corresponding to multiple network slice activations to form a first dataset; acquiring historical network slice performance monitoring data and second decision results corresponding to multiple network slice operation and maintenance to form a second dataset; and training a correlation model based on the first and second datasets. This correlation model characterizes the correlation between network slice quality data and network parameters. Based on this correlation model, automatic activation and / or automatic operation and maintenance of network slices can be achieved. This not only meets users' customized needs and improves the efficiency of network slice activation and / or operation and maintenance, but also optimizes network resources and network configuration strategies, reducing network resource waste. Attached Figure Description

[0051] Figure 1 This is a schematic diagram illustrating the process of enabling network slicing in related technologies;

[0052] Figure 2 This is a schematic diagram illustrating the operational and maintenance methods for network slicing in related technologies.

[0053] Figure 3 This is a flowchart illustrating the network slice management method according to an embodiment of this application;

[0054] Figure 4A and Figure 4B A schematic diagram illustrating the principle of training a BP neural network model;

[0055] Figure 5 This is another flowchart illustrating the network slice management method according to an embodiment of this application;

[0056] Figure 6 This is a flowchart illustrating the automatic network slice management method based on the correlation analysis of network quality and network parameters, as an application embodiment of this application.

[0057] Figure 7 This is a schematic diagram of the network slice management device according to an embodiment of this application;

[0058] Figure 8 This is another structural schematic diagram of the network slice management device according to an embodiment of this application;

[0059] Figure 9 This is a schematic diagram of the network slicing data processing device according to an embodiment of this application.

[0060] Figure 10 This is a schematic diagram of the network slicing management function device according to an embodiment of this application. Detailed Implementation

[0061] The present application will now be described in further detail with reference to the accompanying drawings and embodiments.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0063] In related technologies, in order to realize the management of network slice activation, operation and maintenance, a network slice management system has been added to the 5G network. The network slice management system may include: Communication Service Management Function (CSMF), Network Slice Management Function (NSMF), and Network Slice Subnet Management Function (NSSMF).

[0064] CSMF is mainly used to transform communication service-related requirements into network slicing-related requirements, such as network type, capacity, and QoS (quality of service) requirements, and to send relevant service requirements and management requests to NSMF through an interface with NSMF.

[0065] NSMF is mainly used to receive network slice management requirements related to communication services through the interface with CSMF, manage the lifecycle, performance, and faults of network slices, and can further transform the relevant service requirements into management requirements related to network slice subnets, and send network slice subnet management requirements through the interface with NSSMF.

[0066] NSSMF is primarily used to receive network slice subnet management requests through the interface with NSMF, and to manage the lifecycle, performance, and faults of network slice subnets.

[0067] Network slices and network slice subnets can be managed by the aforementioned network slice management system. A network slice management system can create multiple network slices and multiple network slice subnets, and simultaneously operate and maintain them, i.e., modify network resources and network configurations.

[0068] Figure 1 This illustrates the method for enabling network slicing in related technologies. When enabling network slicing, it typically requires manual intervention on the NSMF to convert the customer's specific service requirements (latency, bandwidth, etc.) into network slicing requirements. These end-to-end network slicing requirements are then further decomposed into network slicing subnet requirements, forming network planning parameters, which are then distributed to each subdomain NSMF for automatic network resource scheduling and parameter configuration. For example... Figure 1 As shown, the methods for enabling network slicing include:

[0069] Step 101: Send customer business requirement information, including a business identifier.

[0070] CSMF sends customer service request information to NSMF, which carries a service identifier.

[0071] Step 102: The customer's business requirements information is manually converted into end-to-end network slicing requirements, and further decomposed into requirements for each network slice subnet, generating network slice instance identifiers.

[0072] On the NSMF side, it is necessary for manual personnel to convert customer business requirements into network slicing requirements based on expert experience, and then decompose them into requirements for each network slice subnet, generating network slice instance identifiers.

[0073] Step 103: NSMF distributes the decomposed network slice subnet requirements (i.e., network planning parameters) to NSMMF, carrying the network slice instance identifier.

[0074] Step 104: Each subdomain NSSMF generates network slice subnet maintenance and management information and network slice subnet instance identifier, and binds the identifier to the network slice instance identifier.

[0075] Step 105: NSSMF passes the virtualized network slice subnet management requirements to the virtualized resource management and orchestration (MANO) function.

[0076] Step 106: MANO allocates virtualization resources, generates network slice subnet resource instances, and configures the resource instances.

[0077] Step 107: MANO sends the virtualized network resource instance generation result to NSMF.

[0078] Step 108: Send the network slice subnet instance generation result, carrying the network slice subnet instance identifier.

[0079] NSSMF sends the network slice subnet instance generation result to NSMF, and the network slice subnet instance generation result carries the network slice subnet instance identifier.

[0080] Step 109: Bind the network slice instance identifier and the network slice subnet instance identifier.

[0081] NSMF binds the network slice instance identifier and the network slice subnet instance identifier.

[0082] Step 110, end-to-end network slicing configuration, including slice service configuration and network connectivity configuration.

[0083] NSMF performs end-to-end network slicing configuration, including slice service configuration and network connectivity configuration.

[0084] Step 111: Return the network slice creation result, carrying the network slice instance identifier.

[0085] NSMF returns the network slice creation result to CSMF. This network slice creation result carries the network slice instance identifier. At this point, the network slice is successfully activated.

[0086] Figure 2 This illustrates a method for network slicing operation and maintenance in related technologies. During service operation, the performance indicators of network slices are monitored. For declining performance indicators, network optimization and adjustment strategies are decided based on human experience. These strategies are then distributed to NSMF via NSMF for network resource scheduling and parameter configuration. Figure 2 As shown, the methods for operating and maintaining network slices include:

[0087] Step 201: Send slice performance indicator monitoring data.

[0088] The PMS (Preventative Maintenance System) periodically sends performance indicator monitoring data to the NSMF.

[0089] Step 202: NSMF compares the performance indicator monitoring data with the service requirement data, and manually decides on network optimization strategies for the declining service-level network performance indicators.

[0090] Here, on the NSMF side, manual decisions on optimization strategies for network slice operation and maintenance need to be made based on expert experience.

[0091] Step 203: Distribute network policies belonging to different subdomains to the corresponding NSMFs, carrying the network slice subnet instance identifier.

[0092] NSMF distributes network policies belonging to different subdomains to the corresponding NSMFs. These network policies carry the identifier of the network slice subnet instance.

[0093] Step 204: Manually decompose the subnet strategy into network element-level resource adjustment strategy and configuration adjustment strategy as needed.

[0094] Here, on the NSSMF side, manual intervention is required based on expert experience to further decompose the subnet strategy into network element-level resource adjustment strategies and configuration adjustment strategies, thereby achieving optimization of network slice subnets.

[0095] In step 205, NSSMF transmits the resource adjustment request to the virtualization resource MANO.

[0096] Step 206: MANO allocates or terminates virtualization resources according to the resource policy, changes the network slice subnet resource instance, and configures the resource instance.

[0097] Step 207: Return the results of the virtualized network resource instance changes.

[0098] MANO returns the changes to the virtualized network resource instance to NSMF.

[0099] Step 208: NSSMF sends the configuration adjustment policy to the Network Element Operation and Maintenance Management Center (OMC).

[0100] Step 209: OMC sends the configuration policy distribution result to NSMF.

[0101] Step 210: NSSMF sends the configuration policy distribution result to NSMF.

[0102] Since both network slice activation and operation / maintenance require manual intervention and expert-based strategy determination, this application provides a network slice management method that automatically enables network slice activation and operation / maintenance in various embodiments.

[0103] like Figure 3 As shown in the figure, this application provides a network slice management method that can be applied to a data analysis system (DAS). The method includes:

[0104] Step 301: Obtain historical service demand data and first decision results corresponding to multiple network slice activations to form the first dataset.

[0105] For example, each time such as Figure 1 After network slicing is activated using the method shown, NSMF can send the dependency data (service requirement data) from manual network design and the results of manual decisions (i.e., the first decision result) to DAS. The service requirement data may include network speed, latency, capacity, network type, etc., and the first decision result can be network element-level network planning parameters. DAS can obtain historical service requirement data and first decision results corresponding to multiple network slice activations to form a first dataset. For example, the first dataset contains no fewer than 300 data entries, with each data sample receiving historical service requirement data as input and the first decision result as output.

[0106] Step 302: Obtain historical network slice performance monitoring data and second decision results corresponding to multiple network slice operation and maintenance to form the second dataset.

[0107] For example, each time such as Figure 2 After the method described is implemented for operation and maintenance, NSMF can send the dependency data (network slice performance monitoring data) for manual optimization strategy selection and the manual decision result (i.e., the second decision result) to DAS. The network slice performance monitoring data may include data such as network speed, latency, and capacity. The second decision result can be a network element-level network optimization strategy, which may include adjustments to network resources and / or network configuration. DAS can obtain historical network slice performance monitoring data and second decision results corresponding to multiple network slice operations and maintenance, forming a second dataset. For example, the second dataset contains no fewer than 700 data entries, with each data sample receiving historical network slice performance monitoring data as input and the second decision result as output.

[0108] Step 303: Train and generate a correlation model based on the first dataset and the second dataset. The correlation model characterizes the correlation between network quality data and network parameters of network slices.

[0109] Understandably, since the correlation model can characterize the correlation between network slice quality data and network parameters, it can be used to automatically enable and / or automatically operate and maintain network slices after enabling them. This can not only meet users' customized needs and improve the efficiency of network slice enabling and / or operation and maintenance, but also optimize network resources and network configuration strategies and reduce network resource waste.

[0110] For example, training and generating a correlation model based on the first dataset and the second dataset includes:

[0111] The first and second datasets are transformed and merged according to the set transformation rules to obtain the merged third dataset;

[0112] The correlation model is obtained by training the third dataset based on a pre-established network model.

[0113] The historical service demand data and the historical network slice performance monitoring data are combined into the network quality data of the third dataset, and the first decision result and the second decision result are combined into the network parameters of the third dataset.

[0114] In one application example, the inputs and outputs of the first dataset are shown in Table 1.

[0115] Table 1

[0116]

[0117]

[0118] In one application example, the inputs and outputs of the second dataset are shown in Table 2.

[0119] Table 2

[0120]

[0121] For example, DAS can transform and merge the first and second datasets to obtain a merged third dataset. For instance, it can first clean outlier data, and then perform transformation according to predefined rules. For example, it can convert enumeration type parameter values ​​to numerical values ​​according to their level, and remove missing data and parameters that cannot be converted to numerical values. Specific transformation rules are not limited here, as long as they meet the data quantification requirements.

[0122] It should be noted that the end-to-end business requirements data provided by the customer primarily concern network performance, while the network slice performance monitoring data mainly focuses on monitoring network performance during operation. For example, network slice performance monitoring data can include resource-level monitoring indicators such as business-level performance metrics and resource utilization rates, as well as complex network environment parameters such as electromagnetic interference. Therefore, there is significant overlap in parameters between the business requirements data and the network slice performance monitoring data. Furthermore, network element-level network optimization strategies are essentially network planning parameters. Consequently, the first and second datasets have a large overlap in both input and output parameters. Therefore, the first and second datasets can be merged to form a third dataset. The input of this third dataset is network quality data, and the output is network parameters.

[0123] It is understandable that the network quality dataset becomes the parameter item of the business requirement data in the first dataset and the parameter item of the network slice performance monitoring data in the second dataset. The network parameters integrate the parameter items of the first decision result and the parameter items of the second decision result in the first dataset.

[0124] For example, the inputs and outputs of the third dataset are shown in Table 3.

[0125] Table 3

[0126] Serial Number Network quality Network parameters 1 downlink speed Core network UPF deployment location 2 Uplink speed Core network UPF isolation level 3 UE speed Transmission isolation level 4 End-to-end delay VPN routing policy 5 Terminal mobility level Wireless RB Reserve Percentage 6 resource utilization ratio Does it offer super uplink?

[0127] It is understandable that when converting data samples from the first or second dataset into data samples from the third dataset, if the input and output parameters of the third dataset do not overlap, the values ​​of the non-overlapping parameter items can be padded with 0.

[0128] It should be noted that the number of data samples in the first dataset is limited by the number of times network slices are enabled, and is often limited, for example, 300. However, network slice performance monitoring data can be statistically analyzed at the minute level. The number of data samples in the second dataset can be larger, for example, no less than 700. By merging the first and second datasets into the third dataset, the number of data samples in the third dataset is enriched, for example, no less than 1,000, which is beneficial to improving the accuracy of the relevance model trained.

[0129] It's important to note that the pre-built network model can be a machine learning model. Machine learning is the core of artificial intelligence, involving many fields such as statistics, system identification, approximation theory, neural networks, optimization theory, computer science, and neuroscience. It studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills, reorganize existing knowledge structures, and continuously improve their performance. Data-driven machine learning is one of the important methods in modern intelligent technology. It starts by looking for patterns in observed data (samples) and uses these patterns to predict future data or data that cannot be observed. Based on the learning method, learning can be divided into: supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Machine learning algorithms include, but are not limited to: logistic regression, hidden Markov methods, support vector machine (SVM) methods, k-nearest neighbor (KNN) methods, three-layer artificial neural network methods, Adaboost algorithm, naive Bayesian classification methods, and decision tree methods, etc. The embodiments of this application do not limit the specific type of machine learning algorithm.

[0130] For example, the pre-established network model is a back propagation (BP) neural network model, and the process of training the third dataset based on the pre-established network model to obtain the correlation model includes:

[0131] The weight values ​​of the BP neural network model are trained bidirectionally based on the network quality data and network parameters of the third dataset to obtain the correlation model.

[0132] Here, the weight values ​​of the bidirectional training BP neural network model include: Training 1 and Training 2. Training 1 refers to training with network quality data as input and network parameters as output, while Training 2 refers to training with network parameters as input and network quality data as output, thereby obtaining a correlation model between network quality and network parameters.

[0133] For example, the specific training steps for Training 1 and Training 2 are as follows:

[0134] Step 1: Pre-set neural network model. Initially, a 3-layer neural network model is built based on the input and output dataset dimensions. The network quality data serves as the input, and the network parameters serve as the output. (The model is shown below.) Figure 4A As shown, the network model with network parameters as input and network quality data as output is as follows: Figure 4B As shown, the filling nodes are coincident parameter nodes.

[0135] Step 2: Randomly initialize weights, intercept term, and learning rate; set node activation function and output layer error function.

[0136] Step 3: Implement forward propagation, including calculating the output values ​​of each node in the hidden layer and the output layer;

[0137] Step 4: Implement backpropagation, including calculating the total error, updating the output layer weights, and updating the hidden layer weights;

[0138] Step 5: Iterate until all data has been trained.

[0139] Understandably, the BP neural network model can learn and store a large number of input-output pattern mappings without needing to reveal the mathematical equations describing these mappings beforehand. Its learning rule uses gradient descent, continuously adjusting the network's weights and thresholds through backpropagation to minimize the sum of squared errors. The topology of the BP neural network model includes an input layer, hidden layers, and an output layer.

[0140] In some embodiments, the method further includes:

[0141] The correlation model is sent to the network slice management function device.

[0142] Understandably, DAS can send the relevance model generated during training to NSMF.

[0143] This application also provides a network slice management method that can be applied to NSMF, such as... Figure 5 As shown, the method includes:

[0144] Step 501: Receive management request data for managing network slices;

[0145] Step 502: Convert the management request data into network quality data of the network slice based on the set conversion rules;

[0146] Step 503: Input the network quality data into the correlation model to obtain the corresponding network parameters;

[0147] Step 504: Manage network slices based on the network parameters.

[0148] Understandably, this correlation model can be used for DAS based on Figure 3 The method shown generates a correlation model. Management request data can correspond to requests to activate network slices or requests to operate and maintain network slices. Thus, based on this correlation model, automatic management of network slice activation and / or operation and maintenance can be achieved.

[0149] For example, the management request data is service requirement data requesting the activation of a network slice; the management of the network slice based on the network parameters includes:

[0150] The network slice is activated based on the network parameters.

[0151] For example, enabling network slicing based on the network parameters includes:

[0152] Based on the network parameters, the required information of the network slice subnet is determined, and a network slice instance identifier is generated;

[0153] The network slice subnet requirement information and the network slice instance identifier are sent to the network slice subnet management function device.

[0154] For example, the management request data is network slice performance monitoring data during the operation of an activated network slice; the management of network slices based on the network parameters includes:

[0155] The network slice is operated and maintained based on the network parameters.

[0156] For example, the operation and maintenance of the network slice based on the network parameters includes:

[0157] Adjust the resources and configuration of the network slice based on the network parameters;

[0158] If it is determined that the adjusted network handover performance monitoring data of the network slice does not meet the target parameters of the SLA, then the network parameters are adjusted based on the connection weights of the correlation model until the network handover performance monitoring data of the network slice meets the target parameters of the SLA.

[0159] The present application will now be described in further detail with reference to an application embodiment.

[0160] like Figure 6As shown in the illustration, this application embodiment provides an automatic network slice management method based on the correlation analysis of network quality and network parameters. This management method consists of three stages: a decision-making capability generation stage, a network slice activation stage, and a network slice operation and maintenance stage (also known as a network slice assurance stage). The following section combines... Figure 6 Each stage will be illustrated with an example.

[0161] like Figure 6 As shown, the decision-making capability generation stage includes:

[0162] Step 601: Network slicing is enabled.

[0163] Step 602: Send the client's end-to-end service requirements and network planning parameters.

[0164] Understandably, after each collaborative interaction between CSMF, NSMF, NSSMF, and MANO to complete network slice activation, NSMF sends the dependency data (business requirement data) from manual network design and the results of manual decision-making (network element-level network planning parameters) to DAS, thus obtaining the network slice activation dataset (i.e., the aforementioned first dataset). The amount of data in this dataset should be no less than 300 records.

[0165] Step 603, Network Slicing Protection.

[0166] Here, network slicing protection refers to the operation and maintenance of already activated network slices.

[0167] Step 604: Send network slice performance monitoring data and network element-level network optimization strategies.

[0168] Understandably, after each collaborative effort between NSMF, NSSMF, MANO, PMS, and OMC to ensure network slicing performance, NSMF sends the dependency data for manual policy selection (network slice performance monitoring data) and the results of manual decision-making (network element-level network optimization strategies) to DAS, thus obtaining the network slice assurance dataset (i.e., the aforementioned second dataset). Since the assurance data can be statistically analyzed at the minute level, the amount of data in this dataset should be no less than 700 records or more.

[0169] Step 605: Generate a correlation model between network quality and network parameters based on the BP neural network model.

[0170] Step 605 may specifically include:

[0171] (1) Data processing.

[0172] Here, the data from the first and second datasets can be cleaned and transformed based on data transformation rules, including outlier handling and normalization. To facilitate subsequent training, the enumeration type parameter values ​​are converted into numerical values ​​according to their level, and missing data and parameters that cannot be converted into numerical values ​​are removed. Since the end-to-end business requirements data proposed by the customer mainly concern network performance, subsequent assurance directly focuses on summarizing, calculating, and monitoring these performance indicators to ensure they are not below customer requirements (the difference between the two is that network slice performance monitoring data includes not only business-level performance indicators but also resource-level monitoring indicators such as resource utilization ratio (load) and complex network environment parameters such as electromagnetic interference). The network element-level network optimization strategy data are essentially network planning parameters. Therefore, the two datasets have a large number of overlapping parameters. These are merged into a single dataset, the network quality and network planning parameter dataset (i.e., the aforementioned third dataset), which can increase the amount of data in the training set (no less than 1000 records).

[0173] (2) Neural network training.

[0174] The weights of the neural network are trained bidirectionally using the BP algorithm (training 1 uses network quality data as input and network planning parameter data as output, and training 2 uses network planning parameter data as input and network quality data as output) to obtain a correlation model between network quality and network planning parameter data.

[0175] Step 606: Send the correlation model and data transformation rules.

[0176] DAS sends the correlation model between the trained network quality and network parameters, along with data transformation rules, to NSMF for subsequent automatic activation and protection of network slices.

[0177] Continue to refer to Figure 6 The decision-making capability generation stage includes:

[0178] Step 607: Send the customer's business requirement data, carrying the business identifier.

[0179] Here, the CSMF carries the network slice service identifier and sends the service requirement data to the NSMF.

[0180] Step 608: NSMF transforms the business requirement data according to the data transformation rules, generates network parameters according to the correlation model, determines the network slice subnet requirements according to the network parameters, and generates network slice instance identifiers.

[0181] Here, NSMF can transform the business requirement data according to the data transformation rules involved in step 605 above to obtain network quality data. The network quality data is then input into the correlation model to output network parameters. Based on these network parameters, the network slice subnet requirements can be determined, and a network slice instance identifier can be generated.

[0182] Step 609: NSMF distributes the network slice subnet requirements to NSMF, carrying the network slice instance identifier.

[0183] Here, the network slice subnet requirements are the network optimization parameters of each network slice subnet. NSMF can send these parameters to NSMF, and each network slice subnet's network optimization parameters carry the network slice instance identifier.

[0184] Step 610: Each subdomain NSSMF generates network slice subnet maintenance and management information and network slice subnet instance identifier, and binds the identifier to the network slice instance identifier.

[0185] Step 611, NSSMF passes the virtualized network slice subnet management requirements to the virtualized resource MANO.

[0186] Step 612: MANO allocates virtualization resources, generates network slice subnet resource instances, and configures the resource instances.

[0187] Step 613: MANO notifies NSMF of the virtualized network resource instance generation result.

[0188] In step 614, NSSMF notifies NSMF of the network slice subnet instance generation result, carrying the network slice subnet instance identifier.

[0189] Step 615: NSMF binds the network slice instance identifier and the network slice subnet instance identifier.

[0190] Step 616: Perform end-to-end network slicing configuration, including service configuration and network connectivity configuration.

[0191] Step 617: NSMF returns the CSMF network slice creation result, carrying the network slice instance identifier.

[0192] Continue to refer to Figure 6 The network slicing protection phase includes:

[0193] Step 618: Send network slice performance monitoring data.

[0194] Understandably, PMS can periodically send the statistical network slice performance monitoring data to NSMF.

[0195] Step 619: Convert the network slice performance monitoring data into network slice network quality data based on the set conversion rules, obtain network parameters according to the correlation model, and perform operation and maintenance on the network slice based on the network parameters.

[0196] For example, NSMF can convert the network slice performance monitoring data according to the data conversion rules involved in step 605 above to obtain network quality data. The network quality data is then input into the correlation model to output network parameters. These network element parameters may include network element-level resource adjustment strategies and configuration adjustment strategies. The network element slices can then be operated and maintained according to these network parameters.

[0197] For example, if policy distribution based on the aforementioned network parameters still fails to bring the actual performance measurement value to the guaranteed target value (e.g., the target parameter of SLA), the network parameters can be adjusted based on the connection weights of the correlation model until the network switching performance monitoring data of the network slice meets the target parameter of SLA.

[0198] For example, the impact of each network parameter on network quality data (connection weight) can be reversed, and the cost of adjusting different strategies can be combined to target the performance indicators that have not reached the guaranteed target value, until the actual performance measurement value reaches the target guaranteed value.

[0199] For example, the selection criteria for this targeted strategy distribution can be calculated as follows:

[0200] Assuming the influence weights of network parameters A, B, C, D, E, and F on network quality parameter G are a1, b1, c1, d1, e1, and f1, respectively, and the cost coefficients for each adjustment level after evaluation are a2, b2, c2, d2, e2, and f2 (the lower the cost, the higher the cost coefficient), then the basis coefficient equals the influence weight multiplied by the cost coefficient, i.e., a1*a2, b1*b2, c1*c2, d1*d2, e1*e2, and f1*f2. The corresponding strategies are then issued sequentially according to the basis coefficients of the network parameters from highest to lowest, adjusting only one level at a time, until the actual performance measurement reaches the target guaranteed value. This ensures that the operational performance of the network slice meets the SLA requirements.

[0201] Step 620: Distribute the network element-level policies belonging to different subdomains to the corresponding NSMFs, carrying the network slice subnet instance identifier.

[0202] Step 621, NSSMF passes the resource adjustment request to the virtualization resource and orchestration function MANO.

[0203] Step 622: MANO allocates virtualization resources, generates network slice subnet resource instances, and configures the resource instances.

[0204] Step 623: MANO notifies NSMF of the virtualized network resource instance generation result.

[0205] Step 624: NSSMF transmits the configuration adjustment policy to the Network Element Operation and Maintenance Management Center (OMC).

[0206] Step 625: OMC notifies NSMF of the policy configuration distribution result.

[0207] Step 626: NSSMF sends the policy configuration distribution result to NSMF.

[0208] It is understood that the method in this application embodiment sends the data (service requirement data, network slice performance monitoring data) on which network design and strategy selection depend during network slice activation and protection, as well as the network design results and strategy selection results from manual analysis and decision-making, to the Data Analysis System (DAS). The DAS then cleans, transforms, merges, and performs correlation analysis on the data to obtain the ability to automatically decide on network planning and design, and automatically decide on network optimization strategies. This method solves the problem of automatically activating network instances that meet customized business needs by combining network resources and configuring network configurations during network slice management, and automatically combining various network strategies during subsequent business operations to continuously ensure the quality of communication services contracted by customers. This improves the efficiency of network slice activation and protection and avoids resource waste.

[0209] In order to implement the method of the embodiments of this application, the embodiments of this application also provide a network slice management device, which corresponds to the network slice management method described above. The steps in the embodiments of the network slice management method described above are also fully applicable to the embodiments of this network slice management device.

[0210] like Figure 7 As shown, the network slice management device includes: a first acquisition module 701, a second acquisition module 702, and a correlation model generation module 703.

[0211] The first acquisition module 701 is used to acquire historical service demand data and first decision results corresponding to multiple network slice activations, forming a first dataset; the second acquisition module 702 is used to acquire historical network slice performance monitoring data and second decision results corresponding to multiple network slice operation and maintenance, forming a second dataset; the correlation model generation module 703 is used to train and generate a correlation model based on the first dataset and the second dataset, wherein the correlation model characterizes the correlation between network slice network quality data and network parameters.

[0212] For example, the correlation model generation module 703 is specifically used for:

[0213] The first and second datasets are transformed and merged according to the set transformation rules to obtain the merged third dataset;

[0214] The correlation model is obtained by training the third dataset based on a pre-established network model.

[0215] The historical service demand data and the historical network slice performance monitoring data are combined into the network quality data of the third dataset, and the first decision result and the second decision result are combined into the network parameters of the third dataset.

[0216] For example, the pre-established network model is a BP neural network model, and the correlation model generation module 703 is specifically used for:

[0217] The weight values ​​of the BP neural network model are trained bidirectionally based on the network quality data and network parameters of the third dataset to obtain the correlation model.

[0218] For example, the network slice management device further includes a sending module 704 for sending the correlation model to the NSMF.

[0219] In practical applications, the first acquisition module 701, the second acquisition module 702, the correlation model generation module 703, and the sending module 704 can be implemented by the processor in the network slice management device. Of course, the processor needs to run the computer program in the memory to implement its functions.

[0220] like Figure 8 As shown, the network slicing management device includes: a receiving module 801, a conversion module 802, a determination module 803, and a management module 804.

[0221] The receiving module 801 is used to receive management request data for managing network slices; the conversion module 802 is used to convert the management request data into network quality data of the network slice based on the set conversion rules; the determining module 803 is used to input the network quality data into the aforementioned correlation model to obtain the corresponding network parameters; and the management module 804 is used to manage the network slices based on the network parameters.

[0222] For example, the management request data is the business requirement data for requesting the activation of network slices, and the management module 804 is specifically used for: activating network slices based on the network parameters.

[0223] For example, the management module 804 enables network slicing based on the network parameters, including:

[0224] Based on the network parameters, the required information of the network slice subnet is determined, and a network slice instance identifier is generated;

[0225] The network slice subnet requirement information and the network slice instance identifier are sent to NSMF.

[0226] For example, the management request data is network slice performance monitoring data during the operation of an activated network slice, and the management module 804 is specifically used for:

[0227] The network slice is operated and maintained based on the network parameters.

[0228] For example, the management module 804 performs operation and maintenance on the network slice based on the network parameters, including:

[0229] Adjust the resources and configuration of the network slice based on the network parameters;

[0230] If it is determined that the adjusted network handover performance monitoring data of the network slice does not meet the target parameters of the Service Level Agreement (SLA), then the network parameters are adjusted based on the connection weights of the correlation model until the network handover performance monitoring data of the network slice meets the target parameters of the SLA.

[0231] In practical applications, the receiving module 801, conversion module 802, determination module 803, and management module 804 can be implemented by the processor in the network slicing management device. Of course, the processor needs to run the computer program in the memory to implement its functions.

[0232] It should be noted that the network slice management device provided in the above embodiments is only illustrated by the division of the above program modules when performing network slice management. In actual 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 network slice management device and the network slice management 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.

[0233] Based on the hardware implementation of the above-described program modules, and in order to implement the method of this application embodiment, this application embodiment also provides a network slice data processing device. This network slice data processing device is equivalent to the aforementioned DAS. Figure 9 This is only an exemplary structure of the network slice data processing device, not the entire structure; it can be implemented as needed. Figure 9 The structure shown may be part or all of the structure.

[0234] like Figure 9As shown, the network slicing data processing device 900 provided in this embodiment includes: at least one processor 901, a memory 902, a user interface 903, and at least one network interface 904. The various components in the network slicing data processing device 900 are coupled together via a bus system 905. It can be understood that the bus system 905 is used to implement communication between these components. In addition to a data bus, the bus system 905 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 9 The general labeled all buses as Bus System 905.

[0235] The user interface 903 may include a monitor, keyboard, mouse, trackball, click wheel, buttons, touchpad, or touch screen.

[0236] The memory 902 in this embodiment is used to store various types of data to support the operation of the network slice data processing device. Examples of such data include any computer program used to operate on the network slice data processing device.

[0237] The network slice management method disclosed in this application can be applied to or implemented by the processor 901. The processor 901 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the network slice management method can be completed by the integrated logic circuits in the hardware of the processor 901 or by instructions in software form. The processor 901 can 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 processor 901 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method 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 can be located in a storage medium, specifically in memory 902. The processor 901 reads information from memory 902 and, in conjunction with its hardware, completes the steps of the network slice management method provided in the embodiments of this application.

[0238] In an exemplary embodiment, the network slicing data processing device may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components to perform the aforementioned methods.

[0239] Based on the hardware implementation of the above program modules, and in order to implement the method of the embodiments of this application, the embodiments of this application also provide a network slice management function device (NSMF). Figure 10 This is merely an exemplary structure of the network slice management device, not the entire structure; implementation is possible as needed. Figure 10 The structure shown may be part or all of the structure.

[0240] like Figure 10 As shown, the network slice management function device 1000 provided in this application embodiment includes: at least one processor 1001, a memory 1002, a user interface 1003, and at least one network interface 1004. The various components in the network slice management function device 1000 are coupled together through a bus system 1005. It can be understood that the bus system 1005 is used to implement communication between these components. In addition to a data bus, the bus system 1005 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 10 The general labeled all buses as Bus System 1005.

[0241] The user interface 1003 may include a monitor, keyboard, mouse, trackball, click wheel, buttons, touchpad, or touch screen.

[0242] The memory 1002 in this embodiment is used to store various types of data to support the operation of the network slice management function device. Examples of such data include any computer program used to operate on the network slice management function device.

[0243] The network slice management method disclosed in this application can be applied to or implemented by the processor 1001. The processor 1001 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the network slice management method can be completed by the integrated logic circuits in the hardware of the processor 1001 or by instructions in software form. The processor 1001 can 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 processor 1001 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method 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 can be located in a storage medium, which is located in the memory 1002. The processor 1001 reads the information in the memory 1002 and, in conjunction with its hardware, completes the steps of the network slice management method provided in the embodiments of this application.

[0244] In an exemplary embodiment, the network slice management function device 1000 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.

[0245] It is understood that memories 902 and 1002 can be volatile or non-volatile memories, or both. Non-volatile memories can be read-only memories (ROM), programmable read-only memories (PROM), erasable programmable read-only memories (EPROM), electrically erasable programmable read-only memories (EEPROM), ferromagnetic random access memories (FRAM), flash memories, magnetic surface memories, optical discs, or compact disc read-only memories (CD-ROM); magnetic surface memories can be disk storage or magnetic tape storage. Volatile memories can be random access memories (RAM), used as external caches. 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.

[0246] 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 memory 902 including a computer program, which can be executed by the processor 901 of a network slice data processing device to complete the steps described in the network slice management method of this application embodiment; or, for example, a memory 1002 including a computer program, which can be executed by the processor 1001 of a network slice management function device to complete the steps described in the network slice management method of this application embodiment. The computer-readable storage medium can be a ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM, etc.

[0247] 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.

[0248] Furthermore, the technical solutions described in the embodiments of this application can be combined arbitrarily without conflict.

[0249] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A network slice management method, characterized by, The method comprises: obtaining historical service demand data and a first decision result corresponding to multiple network slice openings, to form a first data set; obtaining historical network slice performance monitoring data and a second decision result corresponding to multiple network slice running and maintenance, to form a second data set; training a correlation model based on the first data set and the second data set, the correlation model representing the correlation between network quality data and network parameters of a network slice; the training of the correlation model based on the first data set and the second data set comprises: performing conversion and merging processing on the first data set and the second data set based on a set conversion rule, to obtain a merged third data set; training the third data set based on a pre-established network model, to obtain the correlation model; wherein the historical service demand data and the historical network slice performance monitoring data are merged into network quality data of the third data set, and the first decision result and the second decision result are merged into network parameters of the third data set.

2. The method of claim 1, wherein, The pre-established network model is a back propagation (BP) neural network model, and the training of the correlation model based on the pre-established network model comprises: training the weight values of the BP neural network model based on the network quality data and the network parameters of the third data set, to obtain the correlation model.

3. The method of claim 1, wherein, The method further comprises: sending the correlation model to a network slice management function device. 4.A network slice management method, characterized by, The method comprises: receiving management request data for managing a network slice; converting the management request data into network quality data of the network slice based on a set conversion rule; inputting the network quality data into the correlation model generated by the method of any one of claims 1 to 2, to obtain corresponding network parameters; managing the network slice based on the network parameters.

5. The method of claim 4, wherein, The management request data is service demand data for requesting opening of a network slice, and the management of the network slice based on the network parameters comprises: opening the network slice based on the network parameters.

6. The method of claim 5, wherein, The opening of the network slice based on the network parameters comprises: determining demand information of a network slice subnet based on the network parameters, and generating a network slice instance identifier; sending the demand information of the network slice subnet and the network slice instance identifier to a network slice subnet management function device.

7. The method of claim 4, wherein, The management request data is network slice performance monitoring data in a running process of an opened network slice; The management of the network slice based on the network parameters comprises: performing running and maintenance of the network slice based on the network parameters.

8. The method of claim 7, wherein, The performing of the running and maintenance of the network slice based on the network parameters comprises: adjusting resources and configurations of the network slice based on the network parameters; determining that network switching performance monitoring data of the adjusted network slice does not satisfy target parameters of a service level agreement (SLA), and adjusting the network parameters based on connection weights of the correlation model until the network switching performance monitoring data of the network slice satisfies the target parameters of the SLA. 9.A network slice management apparatus, characterized by comprising: The method comprises: The first obtaining module is configured to obtain historical service demand data corresponding to multiple network slice opening and first decision results, and form a first data set. The second obtaining module is configured to obtain historical network slice performance monitoring data corresponding to multiple network slice operation and maintenance and second decision results, and form a second data set. The correlation model generation module is configured to train and generate a correlation model based on the first data set and the second data set, and the correlation model represents the correlation between network quality data and network parameters of a network slice. The correlation model generation module is specifically configured to: perform conversion and merging processing on the first data set and the second data set based on a set conversion rule, and obtain a third data set after merging. The third data set is trained based on a pre-established network model, and the correlation model is obtained. The historical service demand data and the historical network slice performance monitoring data are merged into network quality data of the third data set, and the first decision result and the second decision result are merged into network parameters of the third data set. 10.A network slice management apparatus, characterized by comprising: The method comprises: The receiving module is configured to receive management request data for managing a network slice. The conversion module is configured to convert the management request data into network quality data of the network slice based on a set conversion rule. The determining module is configured to input the network quality data into the correlation model generated by the method of any one of claims 1 to 2 to obtain corresponding network parameters. The management module is configured to manage the network slice based on the network parameters.

11. A network slice data processing device, comprising: The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 8. The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 8. The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 8. 12.A network slice management function device, characterized by comprising: ​ ​ ​ 13. A storage medium having stored thereon a computer program, characterized in that ​

Images