Network element load monitoring method, device, storage medium and electronic device

Abstract

The present disclosure provides a network element load monitoring method, a network element load monitoring device, a computer storage medium and an electronic device, and relates to the field of communication technology. The network element load monitoring method includes obtaining the load of each access point device in the current actual network to obtain the current load; inputting the current load into a pre-trained digital twin load model to obtain the target load of each network element in the current actual network; and scheduling network management resources according to the target load and a preset load threshold. The present disclosure can match the configured resources with the actual demand.

Description

Background Art

[0002] In order to ensure the stable operation of the overall communication system, the network management system can monitor and analyze the business load of each network element in the network, so as to perform business load scheduling when the network element is overloaded, thereby maintaining load balance among the communication nodes.

[0003] Currently, the network management system collects the service load data of the network elements according to a preset period to monitor and analyze the service load of each network element.

[0004] However, the above method requires the network management system to continuously invest fixed network resources and computing resources, which easily leads to a mismatch between the configured resources and actual needs. Summary of the invention

[0005] The present disclosure provides a network element load monitoring method, a network element load monitoring device, a computer storage medium and an electronic device, thereby realizing configuration according to actual demand for resources.

[0006] In the first aspect, an embodiment of the present disclosure provides a network element load monitoring method, which obtains the load of each access point device in the current actual network to obtain the current load; inputs the current load into a pre-trained digital twin load model to obtain the target load of each network element in the current actual network; and schedules network management resources according to the relative size between the target load and a preset load threshold.

[0007] In an optional embodiment of the present disclosure, before inputting the current load into a pre-trained digital twin load model, the network element load monitoring method also includes: obtaining a first historical load data volume of each access point device and a second historical load data volume of each network element to obtain a first load training set containing the first historical load data volume and the second historical load data volume; training the initial digital twin load model based on the first load training set to obtain a target digital twin load model.

[0008] In an optional embodiment of the present disclosure, the actual load of each network element in the current actual network is obtained; the load difference between the actual load and the target load is determined; if the load difference is greater than the preset load difference, the digital twin network element model is corrected based on the actual load and the current load.

[0009] In an optional embodiment of the present disclosure, if the load difference is greater than a preset load difference, the digital twin network element model is corrected, including: if the load difference is greater than the preset load difference, a second load training set is constructed based on the current load of each access point device corresponding to the actual load and the first load training set; and the current digital twin network element model is retrained based on the second load training set until the load difference is less than or equal to the preset load difference.

[0010] In an optional embodiment of the present disclosure, the corresponding relationship between each network element in the actual communication network is obtained to obtain the network element logical topology between each network element; the corresponding relationship between each network element in the network element logical topology is mapped to the digital twin system, and a digital twin network element logical topology for the corresponding relationship between each network element is constructed; according to the digital twin network element logical topology, the load between each network element is scheduled.

[0011] In an optional embodiment of the present disclosure, if the target load is greater than or equal to a preset load threshold, load warning information is generated.

[0012] In an optional embodiment of the present disclosure, if the target load is less than a preset load threshold, the network management resources are recovered from the resource pool.

[0013] In the second aspect, an embodiment of the present disclosure provides a network element load monitoring device, which includes: a data acquisition module for acquiring the load of each access point device in the current actual network to obtain the current load; a data input module for inputting the current load into a pre-trained digital twin load model to obtain the target load of each network element in the current actual network; a resource scheduling module for scheduling network management resources according to the relative size between the target load and a preset load threshold.

[0014] In an optional embodiment of the present disclosure, the data acquisition module is specifically used to obtain the first historical load data volume of each access point device and the second historical load data volume of each network element to obtain a first load training set containing the first historical load data volume and the second historical load data volume; the model training module is used to train the initial digital twin load model based on the first load training set to obtain a target digital twin load model.

[0015] In an optional embodiment of the present disclosure, the data acquisition module is specifically used to obtain the actual load of each network element in the current actual network; the error determination module is used to determine the load difference between the actual load and the target load; and the model correction module is used to correct the digital twin network element model based on the actual load and the current load if the load difference is greater than the preset load difference.

[0016] In an optional embodiment of the present disclosure, the model correction module is used to construct a second load training set based on the current load of each access point device corresponding to the actual load and the first load training set if the load difference is greater than a preset load difference; and retrain the current digital twin network element model based on the second load training set until the load difference is less than or equal to the preset load difference.

[0017] In an optional embodiment of the present disclosure, the data acquisition module is specifically used to obtain the corresponding relationship between each network element in the actual communication network, and construct a network element logical topology between each network element based on the corresponding relationship; the model training module is used to map the corresponding relationship between each network element in the network element logical topology to the digital twin system, and construct a digital twin network element logical topology; the resource scheduling module is used to schedule the load between each network element according to the digital twin network element logical topology.

[0018] In an optional embodiment of the present disclosure, the alarm generation module is used to generate load alarm information if the target load is greater than a preset load threshold.

[0019] In an optional embodiment of the present disclosure, the resource scheduling module is used to reclaim network management resources from the resource pool if the target load is greater than a preset load threshold.

[0020] In a third aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which implements the above network element load monitoring method when executed by a processor.

[0021] In a fourth aspect, an embodiment of the present disclosure provides an electronic device, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the above network element load monitoring method by executing the executable instructions.

[0022] The technical solution disclosed in this disclosure has the following beneficial effects:

[0023] The above-mentioned network element load monitoring method maps each network element in the actual network to the digital twin environment in advance, obtains the network element twin corresponding to each network element, and trains the digital twin load model to obtain the load of each network element twin. The network element load monitoring method obtains the current load by obtaining the load of each access point device in the current actual network; inputs the current load into the digital twin load model to obtain the target load of each network element in the current actual network; and schedules the network management resources according to the relative size between the target load and the preset load threshold. The method predicts the load of each network element in the actual network according to the pre-trained digital twin load model, and determines whether the load is balanced between each network element according to the predicted load and the preset load threshold, so that when the network management system needs to schedule the load of each network element, it configures the corresponding network management resources for it, without releasing the network management resources occupied by each network element when scheduling the load. This method avoids the technical problem in traditional technologies that the network management system needs to continuously occupy network management resources to monitor and schedule the load of each network element, resulting in the mismatch between the configured resources and the actual demand, thereby achieving the technical effect of configuring resources according to the actual demand for resources and improving the flexibility of resource allocation.

[0024] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 The following schematically shows an architecture diagram of a network element load monitoring system in this exemplary embodiment;

[0027] Figure 2 A flow chart of a network element load monitoring method in this exemplary embodiment is schematically shown;

[0028] Figure 3 A schematic diagram of network management resource scheduling in a load balancing situation in this exemplary embodiment is schematically shown;

[0029] Figure 4 A schematic diagram of network management resource scheduling in a load imbalance situation in this exemplary embodiment is schematically shown;

[0030] Figure 5A flowchart of a method for scheduling load between network elements in this exemplary embodiment is schematically shown;

[0031] Figure 6 A schematic diagram schematically shows a network element logical topology expression form of an actual network in this exemplary embodiment;

[0032] Figure 7 A flowchart of another network element load monitoring method in this exemplary embodiment is schematically shown;

[0033] Figure 8 A schematic diagram of the structure of a network element load monitoring device in this exemplary embodiment is schematically shown;

[0034] Fig. 9 A schematic diagram showing the structure of another network element load monitoring device in this exemplary embodiment is shown;

[0035] Fig.10 The structure diagram of an electronic device in this exemplary embodiment is schematically shown. DETAILED DESCRIPTION

[0036] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, exemplary embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the exemplary embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0037] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0038] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the steps. For example, some steps may be decomposed, while some steps may be combined or partially combined, so the actual execution order may change according to the actual situation.

[0039] The network element load monitoring method provided by the exemplary embodiment of the present disclosure can be applied to application scenarios where it is necessary to monitor the load of each network element and schedule network management resources used to adjust the load between each network element according to the load between each network element. Wherein, the network management is a network management system. For example, in a network management system, the network management can use fixed network resources, computing resources and other network management resources to monitor and analyze the load of each network element. When the load of services between network elements is unbalanced, the service load is scheduled to achieve load balancing between network elements.

[0040] In order to avoid overload of network nodes and ensure the stability of the overall communication system, the existing network management system needs to regularly collect the business load of each network element for monitoring and analysis of the network element load and load scheduling operations between network elements.

[0041] Currently, the network management system will continue to invest fixed network resources and computing resources to start the load monitoring process and load scheduling process, monitor and analyze the network element load, and perform load scheduling operations between network elements. However, when the network elements are in a load balancing state, the network management system will still use fixed network resources and computing resources to collect the load of each network element at a preset period to monitor and analyze the load of each network element, resulting in a waste of network management resources.

[0042] On the basis of the above method, the network management system can also adjust the load collection cycle of each network element to an adaptive cycle, and determine the next load collection cycle length in combination with the current load of each network element and the cycle length. However, this method is that the network management system directly collects the load from each network element, and the timeliness and effectiveness of load scheduling cannot be guaranteed. In addition, the network management system still needs to continue to occupy network resources and computing resources, which easily leads to a mismatch between the configured resources and the actual demand, resulting in poor flexibility in resource scheduling.

[0043] Taking the above problems into consideration, the exemplary embodiment of the present disclosure proposes a network element load monitoring method. This method maps each network element in the actual communication network to a digital twin environment, and obtains a pre-trained digital twin load model, and then inputs the load of each access point device in the current actual network into the digital twin load model to obtain the target load of each network element, and finally schedules the network management resources according to the target load and the preset load threshold. This method can avoid the long-term occupation of resources by the load monitoring process, but configures corresponding resources for each network element when it needs network management resources for business scheduling, thereby improving the flexibility of resource configuration. At the same time, this method maps each network element to a digital twin environment, thereby overcoming the delay in network element load monitoring and scheduling operations caused by the network management system periodically collecting load directly from each network element.

[0044] Figure 1 FIG. 1 is an application scenario diagram of a network element load monitoring method provided by an exemplary embodiment of the present disclosure. Figure 1 As shown, this application scenario is implemented based on a network management platform, which includes a network management system and multiple network elements.

[0045] Among them, the network management system can start the task process for each network element, such as load monitoring process, load scheduling process, etc. The network management platform can configure network management resources such as computing resources and network resources for the task process, so that the network management system uses the configured network management resources to perform corresponding tasks in the started task process, such as load monitoring tasks and load scheduling tasks.

[0046] The network management platform can obtain the load of each access point device in the current actual network and obtain the current load. Then the current load is input into the pre-trained digital twin load model to obtain the target load of each network element; finally, various network management resources in the network management system are scheduled according to the target load and the preset load threshold. Among them, the pre-trained digital twin load model maps each network element entity in the actual communication network to the network element twin in the digital twin environment, thereby training a neural network model based on the network element twin.

[0047] It should be understood that Figure 1 In the application scenario diagram shown, the number of network elements and various resource pools is merely exemplary, and greater or lesser numbers fall within the protection scope of this application.

[0048] The following uses the above-mentioned network element management platform as the execution subject and applies the network element load monitoring method to the above-mentioned network element management platform as an example for explanation. Figure 2 The network element load monitoring method provided by the exemplary embodiment of the present disclosure includes the following steps S201 to S203:

[0049] Step S201: Obtain the load of each access point device in the current actual network to obtain the current load.

[0050] The actual network refers to the actual communication service network, and the load refers to the load data of the carried services.

[0051] Exemplarily, the network management platform can obtain the load of each access point device in the current actual network in real time from the data acquisition and control center of the network element.

[0052] Step S202: input the current load into the pre-trained digital twin load model to obtain the target load of each network element.

[0053] Among them, the digital twin load model is a neural network model based on digital twin technology, which maps each network element in the actual communication service network to the digital twin environment to form a network element twin, so as to obtain the business load of each network element twin.

[0054] Exemplarily, the input data of the digital twin load model is the load of each access point device, and the output data is the target load of each network element.

[0055] Step S203: Schedule network management resources according to the relative size between the target load and the preset load threshold.

[0056] The network management resources are computing resources and network resources used for monitoring and scheduling the load of each network element. The network management can start the load monitoring process and load scheduling process for each network element to perform monitoring, analysis, scheduling and other operations on the network element load.

[0057] Exemplarily, after obtaining the target load of each network element in step S202, the target load of each network element can be compared with the preset load threshold of each network element to determine whether the load of each network element is balanced. If the load of each network element is balanced, the load monitoring process and the load scheduling process are dormant, and the network management resources such as computing resources and network resources occupied by the network management system are released so that the network element management platform can perform elastic calls. If the load of each network element is unbalanced, the network management wakes up the load monitoring process and the load scheduling process, and recovers the required computing resources and network resources from the resource pool.

[0058] In the technical solutions provided by some embodiments of the present disclosure, each network element in the actual network is mapped to a digital twin environment in advance, the network element twin corresponding to each network element is obtained, and the digital twin load model is trained to obtain the load of each network element twin. The network element load monitoring method obtains the current load by obtaining the load of each access point device in the current actual network; inputs the current load into the digital twin load model to obtain the target load of each network element in the current actual network; and schedules the network management resources according to the relative size between the target load and the preset load threshold. The method predicts the load of each network element in the actual network according to the pre-trained digital twin load model, and determines whether the load is balanced between each network element according to the predicted load and the preset load threshold, so that when the network management system needs to schedule the load of each network element, it configures the corresponding network management resources for it, without releasing the network management resources occupied by each network element when scheduling the load of each network element. This method avoids the technical problem in traditional technologies that the network management system needs to continuously occupy network management resources to monitor and schedule the load of each network element, resulting in the mismatch between the configured resources and the actual demand, thereby achieving the technical effect of configuring resources according to the actual demand for resources and improving the flexibility of resource allocation.

[0059] In an exemplary embodiment of the present disclosure, it is possible to determine whether the loads among network elements are balanced based on the target load and a preset load threshold, thereby scheduling network management resources based on the determination result.

[0060] According to some embodiments of the present disclosure, if the target load is greater than a preset load threshold, the network management resources are recovered from the resource pool. If the target load is less than or equal to the preset load threshold, the occupied network management resources are released.

[0061] If the target load is greater than the preset load threshold, it is determined that the loads among the network elements are unbalanced; if the target load is less than or equal to the preset load threshold, it is determined that the loads among the network elements are balanced.

[0062] The following will be combined Figure 3 , Figure 4 The network management resource scheduling process under load balancing and load imbalance conditions is schematically illustrated respectively.

[0063] Figure 3 The following schematically shows a schematic diagram of network management resource scheduling in a load balancing situation in this exemplary embodiment. Figure 3 The network management data center contains various resource pools, which include network management resources such as computing resources and network resources, and each network element shares the network management resources with each tenant.

[0064] When the network elements are load balanced, the network manager suspends the load monitoring task and load scheduling task of each network element, that is, the load monitoring process and load scheduling process of each network element are put to sleep, thereby releasing the resources occupied by the network manager so that the network element management platform can schedule the released network manager resources to execute the task requests initiated by each tenant.

[0065] Figure 4 The following schematic diagram shows a schematic diagram of network management resource scheduling in a load imbalance situation in this exemplary embodiment. Figure 4 Similarly, the network management data center contains various resource pools, which include network management resources such as computing resources and network resources, and each network element shares the network management resources with each tenant.

[0066] When the loads among network elements are unbalanced, the load monitoring process and load scheduling process of each network element are awakened, and the network management starts to execute the load monitoring and load scheduling tasks of each network element, and then reclaims the required network management resources such as computing resources and network resources from the resource pool. That is, the resources occupied by tenant tasks are released so that the fixed network management resources can be reclaimed to execute load monitoring and load scheduling tasks until the network elements are in a load-balanced state again.

[0067] In an exemplary embodiment of the present disclosure, when network management resources are recovered, the tasks being executed by the required computing resources and network resources can be stopped and scheduled to other resources, so that the required network management resources are idle for load scheduling operations between network elements, while real-time monitoring is performed until the load between network elements is balanced.

[0068] Exemplarily, when implementing the load scheduling of each network element, the scheduling can be based on the order of service and tenant priority in each network element from high to low. For example, in mobile communication services, there are different package levels for providing different bandwidth rate services to tenants. In the cloud private network, A-level, AA-level, and AAA-level dedicated line guarantee services are provided, and it is agreed that high-level services and tenants are prioritized to stay, thereby reducing network fluctuations and scheduling service network elements for low-level services and users.

[0069] When the target load is less than the preset load threshold, the corresponding network management resources required are recovered from the resource pool, so that resources can be configured when the network management needs to perform load scheduling, and resources can be released when load scheduling is not required, so as to match the configured resources with the actual needs and improve the flexibility of resource configuration. At the same time, it can avoid the waste of resources caused by the long-term occupation of resources by the load monitoring process and load scheduling process of the network management in traditional technologies, and improve the utilization rate of resources.

[0070] According to some other embodiments of the present disclosure, if the target load is greater than a preset load threshold, load warning information is generated.

[0071] Exemplarily, when the target load is less than or equal to the preset load threshold, the digital twin load model does not issue a load warning message, at which time the load monitoring process and load scheduling process of the network management are dormant, and the load monitoring task and load scheduling task are suspended. When the target service load is greater than the preset load threshold, the digital twin load model issues a load warning message, and after receiving the load warning message, the network management sends a notification message to the network element management platform so that the network element management platform can reclaim network resources and computing resources and other network management resources, and execute the load monitoring and analysis tasks of each network element and the network element load scheduling tasks.

[0072] By generating load warning information when the target business load is greater than or equal to the preset load threshold, it can be used to prompt the network manager to schedule resources, thereby facilitating the network element management platform to configure the resources required by the network manager, thereby improving resource utilization and resource configuration flexibility.

[0073] In an exemplary embodiment of the present disclosure, when executing load scheduling between network elements, the corresponding relationship between the network elements in the actual communication network can be obtained, and a network element logical topology between the network elements can be constructed based on the corresponding relationship; the corresponding relationship between the network elements in the network element logical topology is mapped to the digital twin system, and a digital twin network element logical topology for the corresponding relationship between the network elements is constructed; and the load between the network elements is scheduled according to the digital twin network element logical topology.

[0074] Based on digital twin technology, each network element in the actual network is mapped to the digital twin environment to form a network element twin.

[0075] Exemplarily, when performing load scheduling between network elements, scheduling can be performed according to the digital twin network element logical topology between each network element twin in the digital twin environment, and the digital twin network element logical topology shows the corresponding relationship between each network element twin. For example, for network elements A, B, and C, network element A can perform load scheduling to network element B, and network element A cannot perform load scheduling to network element C. When starting load scheduling for network element A, it can only be routed to network element B.

[0076] Figure 5 A flowchart of a method for scheduling load between network elements in this exemplary embodiment is schematically shown. Figure 5 In an optional embodiment of the present disclosure, the process of scheduling the load between network elements includes at least steps S501 to S503:

[0077] Step S501: Acquire network element logical topology data of an actual network.

[0078] The network element logical data is used to describe the connection relationship between network elements in the actual communication network.

[0079] Step S502: construct a logical topology of network elements of the actual network.

[0080] The network element logical topology is constructed by using the network element logical data of the actual network.

[0081] Exemplarily, a directed graph expression may be used to illustrate the connection relationship between network elements in an actual network, so as to construct a logical topology of network elements in the actual network.

[0082] Figure 6 The following schematic diagram shows the logical topology expression of a network element in an actual network in this exemplary embodiment. Let D(V, E) be a directed graph, where V = {v1, v2, v3, ..., vn} is each network element node in the actual network, and E = {e1, e2, e3, ..., em} is the connection relationship between each network element in the actual network. Figure 4 As shown, taking n=6, m=9 as an example, in the directed graph D(V, E), the network element nodes V={v1, v2, v3, ..., v6} in the actual network, and the connection relationship E={e1, e2, e3, ..., e9} of the network elements in the actual network.

[0083] The logical topology of network elements in an actual network can be represented by the adjacency matrix A(D) of the directed graph D(V, E), wherein the logical topology of network elements in an actual network is represented by the adjacency matrix A(D), and the adjacency matrix A(D) can also be understood as the routing probability matrix between network elements. The routing probability matrix is ​​expressed as shown in formula (1):

[0084] A(D)=(aij) nxm (1)

[0085] Among them, the adjacency matrix A(D) is the routing probability matrix between each network element, aij represents the routing probability starting from the network element node vi and ending at the network element node vj, D is a directed graph, n is the number of network element nodes in the actual network, and m is the connection relationship between each network element in the actual network, that is, the routing path between each network element.

[0086] Step S503: Map the network element logical topology of the actual network to the digital twin environment to obtain the digital twin network element logical topology.

[0087] Building a digital twin network element logical topology is to map the network elements in the actual network and the connection relationships between them into the digital twin environment to simulate the actual network environment online. This enables online data acquisition and simulation operations, avoiding frequent calls to various computing resources in the actual communication system to perform various operations.

[0088] After building the logical topology of the digital twin network element, when executing load scheduling between network elements, scheduling can be performed according to the scheduling routing probability in the logical topology of the digital twin network element. Figure 6 Taking the network element node v5 in the digital twin network element logical topology as an example, v5 can migrate the load outward to v1 and v4, and can also receive the load migrated inward by v6 and v2. When the load of network element v5 exceeds the preset load threshold, the load can be migrated outward to v1 and v4 respectively with a preset probability.

[0089] It should be understood that the above-mentioned preset probability can be adaptively adjusted according to the load conditions of network elements v1 and v4. For example, if the available capacity of network element v1 is greater than the available capacity of network element v4, the probability of load migration from network element v5 to network element v1 is greater than the probability of load migration to network element v4.

[0090] By mapping the logical topology of network elements in the actual network to the digital twin environment, the routing process of network elements in the actual network can be mapped online for simulation, avoiding system delays caused by traditional technologies directly performing load scheduling on the logical topology of each network element in the actual network, thereby improving the timeliness and efficiency of load scheduling between network elements.

[0091] In an exemplary embodiment of the present disclosure, the first historical load data volume of each access point device and the second historical load data volume of each network element are obtained to obtain a first load training set; the initial digital twin load model is trained based on the first load training set to obtain a target digital twin load model.

[0092] Among them, the first amount of historical load data may be the load status of each network element, the performance status of each network element, etc.; the second amount of historical load data may be the load status of each access point, etc.; the initial digital twin load model may be a long short-term memory model (Long Short Term Memory, LSTM).

[0093] Exemplarily, after obtaining the first historical load data of each network element and the second historical load data of each access point, the data may be preprocessed, wherein the first historical load data and the second historical load data are the absolute values ​​of the loads of the services carried by each network element.

[0094] Assume that (c1, c2, c3, c4, ..., l1, l2, l3, l4, ...) represents the load of each network element at a certain historical moment, where ci represents the load of each access point device at that moment, and li represents the load of each network element at that moment. Then the load data of multiple historical moments can be represented in the form of an MxN matrix.

[0095] According to some embodiments of the present disclosure, the MxN matrix data of the above-mentioned multiple historical moments are preprocessed to obtain a source data format in the form of an MxN matrix as shown in formula (2).

[0096]

[0097] In formula (2), c mj represents the load of the jth access point device at the mth time, l mj Indicates the load of the jth network element at the mth time.

[0098] The source data format shown in the above formula (2) can be used as the first load training set for training the target digital twin load model.

[0099] In an exemplary embodiment of the present disclosure, when preprocessing the load data in the form of an MxN matrix at multiple historical moments, the load data may be subjected to logic checks, integrity checks, correction of missing values, spelling errors, data cleaning, deduplication, etc., so as to convert the load data into a standard form for model training. At the same time, if the load ratio is to be used as the first load training set, the data may be standardized to improve the subsequent usability of the data.

[0100] Exemplarily, based on the digital twin network element logical topology obtained in step S302, a digital twin network element load model based on a long short-term memory model LSTM is constructed through a first load training set. LSTM can also be called a time recurrent neural network model.

[0101] According to some embodiments of the present disclosure, considering that the first historical load data of each access point device and the second historical load data of each network element collected in the first load training set have time characteristics, and the load of some network elements is continuous and long-lasting, the time recurrent neural network LSTM can be used to train the digital twin load model.

[0102] For example, f t It is used to output the changes of the corresponding existing network element load, and can control the load changes of the network element load model. t The expression of is as follows:

[0103] f t =σ(W f ·[h t-1 ,x t ]+b f ) (3)

[0104] According to the LSTM network: t Represents the coefficient matrix of the forget gate link of the LSTM network, W f Represents a vector [h t-1,x t The product of the input dimension of ] and the output dimension of the hidden layer of the LSTM network, b f Represents the output dimension of the hidden layer of the LSTM network.

[0105] The input gate in LSTM corresponds to the newly added load retention matrix i of each network element t , add load retention matrix i t It is used to determine the degree of retention of newly added network element load data. Specifically, it is shown in formula (4):

[0106] i t =σ(W i ·[h t-1 ,x t ]+b i ) (4)

[0107] According to the LSTM network: t represents the newly added load reservation matrix, W i Represents a vector [h t-1 ,x t The product of the input dimension of ] and the output dimension of the hidden layer of the LSTM network, b i Represents the output dimension of the hidden layer of the LSTM network.

[0108] Added load input matrix C t ~: (According to the update gate layer specified by the LSTM network, it represents the information matrix brought to the LSTM network by the newly added network element load data)

[0109] C t ~=tanh(W c ·[h t-1 ,x t ]+b c ) (5)

[0110] In formula (5), C t ~ represents the newly added load input matrix, and tanh represents the hyperbolic tangent function, which is used as an activation function to normalize the load input content. c Represents a vector [h t-1 ,x t The product of the input dimension of ] and the output dimension of the hidden layer of the LSTM network. c Represents the output dimension of the hidden layer of the LSTM network.

[0111] Based on the above formulas (3), (4) and (5), C t Represents the network element load state matrix at time t: (Retains the load information of the previous moment and retains the parameter matrix for the newly added load.) C t As shown in formula (6):

[0112] C t =f t *C t-1 +i t *C t ～ (6)

[0113] The model parameters of the digital twin network element model are shown in formula (7) and formula (8):

[0114] o t =σ(W o ·[h t-1 ,x t ]+b o ) (7)

[0115] h t =o t *tanh(C t ) (8)

[0116] In formula (7) and formula (8), o t Represents the state output matrix of the output gate layer, W o Represents a vector [h t-1 ,x t The product of the input dimension of ] and the output dimension of the hidden layer of the LSTM network, b o Represents the LSTM network hidden layer output dimension, LSTM model hidden layer state h t It corresponds to the network element performance status at time t.

[0117] In the above formulas (3) to (8), the hidden state h of LSTM is t-1 Corresponding network element performance status; input vector x t corresponds to the state matrix of each access point at the current moment; the cell state corresponds to the load state of each network element, σ represents the sigmoid function, and its output is between 0 and 1.

[0118] By constructing a first load training set through the first historical load data of each access point device in the actual network and the second historical load data of each network element to train the digital twin load model in the digital twin environment, it can avoid the system delay caused by traditional technology directly collecting load from each network element in the actual network and the untimely technical problems of load scheduling, thereby improving the load scheduling efficiency.

[0119] After training the digital twin load model, the load of each network element in the actual network can be predicted. That is, the current load of each access point collected in real time is input into the digital twin load model to obtain the target load of each network element in the current actual network.

[0120] In order to improve the accuracy of the digital twin network element model in judging the load balancing between network elements, or the accuracy of generating load warning information, the network element management platform can regularly revise the digital twin network element model.

[0121] In an exemplary embodiment of the present disclosure, the actual business load of each network element in the current actual network is obtained; the load difference between the actual business load and the target business load is determined; if the load difference is greater than the preset load difference, the digital twin network element model is corrected.

[0122] Exemplarily, the network element management platform can obtain the actual business load of each network element in the current actual network at a preset period to verify the target business load predicted by the digital twin network element model, thereby correcting the digital twin network element model when the load difference is greater than the preset load difference.

[0123] According to some embodiments of the present disclosure, the load difference may be determined by calculating a performance indicator of the digital twin network element model. The performance indicator may be a relative measure of variance or mean square error.

[0124] The variance of the load data is calculated as shown in formula (9):

[0125]

[0126] In formula (9), R 2 is the variance of the load data, SS res is the residual sum of squares, SS tot is the total residual sum of squares.

[0127] The mean square error of the load data is calculated as shown in formula (10):

[0128]

[0129] In formula (10), MSE is the mean square error of the load data, n represents the number of load data samples, and Y i represents the i-th load sample vector, represents the mean vector.

[0130] By calculating the variance and mean square error of the load data of each network element, the possibility of errors can be analyzed. If the calculated load difference is greater than the preset load difference, the model parameters of the digital twin network element model need to be optimized and adjusted.

[0131] In an exemplary embodiment of the present disclosure, if the load difference is greater than a preset load difference, a second load training set is constructed based on the current load of each access point device corresponding to the actual business load and the first load training set; the current digital twin network element model is retrained based on the second load training set until the load difference is less than or equal to the preset load difference.

[0132] Exemplarily, when optimizing and adjusting the model parameters of the digital twin network element model, the actual business load of each network element in the current actual network, the actual business load of each corresponding access point device, and the first historical load data of each access point device in the first load training set and the second historical load data of each network element can be used to reconstruct a new training set to retrain the current digital twin network element model.

[0133] By adding the current load of each access point device with detected load error and the current load of each network element to the first load training set to form a second load training set, the model parameters of the digital twin network element model can be better corrected to improve the accuracy of the digital twin network element model in predicting the load of each network element, thereby timely tracking the load of each network element and realizing timely and efficient load migration.

[0134] The following will refer to Figure 7 The entire process of the network element load monitoring method according to the exemplary embodiment of the present disclosure is described in detail.

[0135] First, according to some embodiments of the present disclosure, it is necessary to pre-train the digital twin network element model. In step S701, the network element logical topology data of the actual network is obtained. The network element logical topology data is the connection relationship between each network element in the actual network.

[0136] In step S702, a digital twin network element logical topology is constructed in a digital twin environment.

[0137] Exemplarily, each network element node and the connection relationship of each network element of the actual network are represented in a directed graph to determine the network element logical topology in the actual network. The network element logical topology in the actual network is mapped to the digital twin environment to construct the digital twin network element logical topology in the digital twin environment.

[0138] In step S703, a first load training set is constructed.

[0139] Exemplarily, historical load data of each network element and historical load data of each access point device at multiple time points may be acquired, and the data may be preprocessed to construct a standardized first load training set.

[0140] After obtaining the digital twin network element logical topology and the first load training set, step S704 can be executed to train the digital twin network element model based on LSTM to obtain the digital twin network element model.

[0141] Exemplarily, considering that the collected load data of each network element and the load data of each access point device have time characteristics, and the load data of some network elements are continuous and long-lasting, LSTM is used for model training.

[0142] Then, in step S705, the load of each access point device in the current actual network is collected and input into the digital twin network element model. Exemplarily, the load of each access point device is collected to input into the trained digital twin network element model, thereby obtaining the target load of each network element in the current actual network.

[0143] Finally, step S706 is executed to determine whether a load warning signal is generated.

[0144] Exemplarily, the digital twin network element model can determine whether the loads are balanced among the network elements based on the target loads of the network elements, that is, it can determine by comparing the target load with the preset load threshold. If the target load is greater than the preset load threshold, it is determined that the loads are not balanced among the network elements, and a load warning message is generated; if the target load is less than or equal to the preset load threshold, it is determined that the loads are balanced among the network elements, and no load warning message is generated.

[0145] If no load warning signal is generated, step S707 is executed to release the network management resources, that is, the load monitoring process and load scheduling process of the network management are dormant, and the load monitoring task and load scheduling task are suspended.

[0146] On the contrary, if a load warning signal is generated, step S708 is executed to recycle network management resources. After receiving the load warning information, the network management sends a notification message to the network element management platform so that the network element management platform can recycle network resources such as network resources and computing resources, and perform load monitoring and analysis tasks of each network element and network element load scheduling tasks.

[0147] After the network management resources are recovered, step S709 can be executed in the load scheduling process to perform load scheduling according to a preset scheduling scheme. The preset scheduling scheme can be one or more.

[0148] During any of the above steps, the current digital twin network element model can be corrected in real time to improve the accuracy and timeliness of the early warning information generated by the digital twin network element model.

[0149] In order to implement the above network element load monitoring method, a network element load monitoring device is provided in an embodiment of the present disclosure. Figure 8A schematic architecture diagram of a network element load monitoring device is shown.

[0150] The network element load monitoring device 800 includes a data acquisition module 801 , a data input module 802 and a resource scheduling module 803 .

[0151] The data acquisition module 801 is used to obtain the load of each access point device in the current actual network to obtain the current load; the data input module 802 is used to input the current load into the pre-trained digital twin load model to obtain the target load of each network element in the current actual network; the resource scheduling module 803 is used to schedule network management resources according to the target load and the preset load threshold.

[0152] The network element load monitoring device 800 provided in the embodiment of the present disclosure can execute the technical solution of the network element load monitoring method in any of the above-mentioned embodiments. Its implementation principle and beneficial effects are similar to the implementation principle and beneficial effects of the network element load monitoring method. Please refer to the implementation principle and beneficial effects of the network element load monitoring method, which will not be repeated here.

[0153] Furthermore, another network element load monitoring device 900 is provided in an embodiment of the present disclosure. Fig. 9 A schematic architecture diagram of another network element load monitoring device 900 is schematically shown.

[0154] The network element load monitoring device 900 includes a data acquisition module 901 , a model training module 902 , an error determination module 903 , a model correction module 904 , an alarm generation module 905 , and a resource scheduling module 906 .

[0155] In an optional embodiment, the data acquisition module 901 is specifically used to obtain the first historical load data volume of each access point device and the second historical load data volume of each network element to obtain a first load training set containing the first historical load data volume and the second historical load data volume; the model training module 902 is used to train the initial digital twin load model based on the first load training set to obtain a target digital twin load model.

[0156] In an optional embodiment, the data acquisition module 901 is specifically used to obtain the actual business load of each network element in the current actual network; the error determination module 903 is used to determine the load difference between the actual business load and the target business load; the model correction module 904 is used to correct the digital twin network element model if the load difference is greater than the preset load difference.

[0157] In an optional embodiment, the model correction module 904 is used to construct a second load training set based on the current load of each access point device corresponding to the actual business load and the first load training set if the load difference is greater than the preset load difference; and retrain the current digital twin network element model based on the second load training set until the load difference is less than or equal to the preset load difference.

[0158] In an optional embodiment, the data acquisition module 901 is specifically used to obtain the connection relationship between each network element in the actual communication network to obtain the network element logical topology between each network element; the model training module 902 is used to map the network element logical topology to the digital twin system to construct the digital twin network element logical topology; the resource scheduling module 906 is used to schedule the load between each network element through network management resources according to the digital twin network element logical topology

[0159] In an optional embodiment, the alarm generation module 905 is used to generate load alarm information if the target service load is greater than a preset load threshold.

[0160] In an optional embodiment, the resource scheduling module 906 is configured to reclaim network management resources from the resource pool if the target load is greater than a preset load threshold.

[0161] The network element load monitoring device 900 provided in the embodiment of the present disclosure can execute the technical solution of the network element load monitoring method in any of the above-mentioned embodiments. Its implementation principle and beneficial effects are similar to the implementation principle and beneficial effects of the network element load monitoring method. Please refer to the implementation principle and beneficial effects of the network element load monitoring method, which will not be repeated here.

[0162] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present invention may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of the present specification.

[0163] The program product for implementing the above method according to an embodiment of the present invention may adopt a portable compact disk read-only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus, or device.

[0164] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0165] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0166] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0167] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0168] In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0169] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0170] Refer to the following Fig.10 The electronic device 1000 according to this embodiment of the present invention will be described. Fig.10 The electronic device 1000 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0171] like Fig.10 As shown, the electronic device 1000 is in the form of a general computing device. The components of the electronic device 1000 may include but are not limited to: the at least one processing unit 1010, the at least one storage unit 1020, a bus 1030 connecting different system components (including the storage unit 1020 and the processing unit 1010), and a display unit 1040.

[0172] The storage unit stores program codes, which can be executed by the processing unit 1010, so that the processing unit 1010 performs the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit 1010 can perform the following steps: Figure 2 Steps S201 to S203 shown in FIG.

[0173] The storage unit 1020 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 10201 and / or a cache storage unit 10202 , and may further include a read-only storage unit (ROM) 10203 .

[0174] The storage unit 1020 may also include a program / utility 10204 having a set (at least one) of program modules 10205, such program modules 10205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0175] Bus 1030 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0176] The electronic device 1000 may also communicate with one or more external devices 1100 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1000, and / or may communicate with any device that enables the electronic device 1000 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1050. Furthermore, the electronic device 1000 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 1060. As shown, the network adapter 1060 communicates with other modules of the electronic device 1000 via a bus 1030. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1000, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0177] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0178] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0179] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0180] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and examples are to be considered as exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0181] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A network element load monitoring method, characterized in that: include: Obtain the load of each access point device in the current actual network to obtain the current load; The current load is input into a pre-trained digital twin load model to obtain a target load of each network element in the current actual network, wherein the pre-trained digital twin load model is obtained by training in a digital twin environment through a first load training set constructed by a first historical load data volume of each access point device in the actual network and a second historical load data volume of each network element; The current load is input into the pre-trained digital twin load model to obtain the target load of each network element in the current actual network, including: The current load is input into a pre-trained digital twin load model to obtain the service load of each corresponding network element twin in the digital twin environment based on the digital twin load model; wherein each network element twin is formed by mapping each access point device in the actual network to the digital twin environment based on the digital twin technology; According to the service load of each network element twin and the digital twin load model, the load of each network element in the actual network is predicted to obtain the target load of each network element in the current actual network; The network management resources are scheduled according to the relative size between the target load and a preset load threshold.

2. The network element load monitoring method according to claim 1, characterized in that: The network element load monitoring method further includes: The initial digital twin load model is trained based on the first load training set to obtain a target digital twin load model, wherein the initial digital twin load model is a long short-term memory model (LSTM).

3. The network element load monitoring method according to claim 1, characterized in that: Also includes: Obtaining the actual load of each network element in the current actual network; Determining a load difference between the actual load and the target load; If the load difference is greater than a preset load difference, the digital twin network element model is corrected based on the actual load and the current load.

4. The network element load monitoring method according to claim 3, characterized in that: If the load difference is greater than a preset load difference, the digital twin network element model is corrected, including: If the load difference is greater than a preset load difference, constructing a second load training set based on the current load of each access point device corresponding to the actual load and the first load training set; The current digital twin network element model is retrained based on the second load training set until the load difference is less than or equal to the preset load difference.

5. The network element load monitoring method according to claim 1, characterized in that: Also includes: Acquire the corresponding relationship between the network elements in the actual communication network, and construct the network element logical topology between the network elements based on the corresponding relationship; Mapping the corresponding relationship between each of the network elements in the network element logical topology to the digital twin system, and constructing a digital twin network element logical topology for the corresponding relationship between each of the network elements; According to the digital twin network element logical topology, the load between the network elements is scheduled.

6. The network element load monitoring method according to claim 1, characterized in that: Also includes: If the target load is greater than a preset load threshold, a load warning message is generated.

7. The network element load monitoring method according to claim 1, characterized in that: The network management resources are scheduled according to the target load and the preset load threshold, including: If the target load is greater than the preset load threshold, the network management resource is recovered from the resource pool.

8. A network element load monitoring device, characterized in that: include: The data acquisition module is used to obtain the load of each access point device in the current actual network and obtain the current load; A data input module is used to input the current load into a pre-trained digital twin load model to obtain a target load of each network element in the current actual network, wherein the pre-trained digital twin load model is obtained by training in a digital twin environment through a first load training set constructed by a first historical load data volume of each access point device in the actual network and a second historical load data volume of each network element; Inputting the current load into a pre-trained digital twin load model to obtain a target load of each network element in the current actual network, including: inputting the current load into a pre-trained digital twin load model to obtain the business load of each network element twin corresponding to the digital twin environment based on the digital twin load model; wherein each network element twin is formed by mapping each access point device in the actual network to the digital twin environment based on digital twin technology; predicting the load of each network element in the actual network according to the business load of each network element twin and the digital twin load model to obtain the target load of each network element in the current actual network; The resource scheduling module is used to schedule network management resources according to the relative size between the target load and a preset load threshold.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the network element load monitoring method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; The processor is configured to execute the network element load monitoring method according to any one of claims 1 to 7 by executing the executable instructions.

Images