Internet leased line speed regulation method and device, electronic equipment and medium

Abstract

The application provides an Internet leased line speed regulation method and device, electronic equipment and medium. The method comprises the following steps: acquiring line information of a current Internet leased line, wherein the line information comprises time information, bandwidth utilization and network quality parameters at a current time; inputting the line information of the Internet leased line into a time sequence-based prediction model to obtain bandwidth utilization of the Internet leased line at a next time output by the prediction model; wherein the prediction model is a pre-trained model; and performing leased line speed regulation on the Internet leased line according to the bandwidth utilization of the Internet leased line at the next time and a bandwidth utilization threshold of the Internet leased line at the next time. The method solves the problem that bandwidth resources cannot be fully utilized in traditional Internet leased line services.

Description

Technical Field

[0001] This application relates to communication technology, and more particularly to a method, apparatus, electronic device, and medium for adjusting the speed of a leased Internet line. Background Technology

[0002] In recent years, with the emergence of new technologies such as 5G, traditional home broadband services have become saturated, leading to new changes in the services carried by traditional metropolitan area networks. Internet leased line services have gradually emerged, creating new demands. Internet leased line services refer to services that connect customers' networks and devices to the Internet via dedicated links, providing customers with real-time online internet access at various speeds.

[0003] Traditional internet leased lines provide users with fixed bandwidth dedicated lines, but during peak business periods, line quality often deteriorates, affecting user experience; while on holidays or at specific times, the lines are idle, increasing the cost of subscribing to leased lines. Therefore, traditional internet leased line services suffer from the problem of not fully utilizing bandwidth resources. Summary of the Invention

[0004] This application provides a method, apparatus, electronic device, and medium for adjusting the speed of a leased Internet line, in order to solve the problem that traditional leased Internet line services cannot fully utilize bandwidth resources.

[0005] On the one hand, this application provides a method for adjusting the speed of a dedicated internet line, including:

[0006] Obtain the current internet leased line information, which includes the current time information, bandwidth utilization, and network quality parameters;

[0007] The current internet leased line information is input into a time-series-based prediction model to obtain the bandwidth utilization rate of the internet leased line at the next time step, as output by the prediction model; wherein, the prediction model is a pre-trained model.

[0008] Based on the bandwidth utilization rate of the internet leased line at the next time moment and the bandwidth utilization rate threshold of the internet leased line at the next time moment, the internet leased line speed is adjusted.

[0009] In one possible implementation, the method further includes:

[0010] During the predetermined data collection time, line information is collected according to a predetermined cycle as feature parameters for model training. The feature parameters include time information, bandwidth utilization, and network quality parameters.

[0011] The time information in the feature parameters is one-hot encoded and concatenated with the corresponding feature parameters to obtain the model training vector;

[0012] An initial model is established, which is a time-series-based recurrent neural network model, including a feature extraction layer and a fully connected layer;

[0013] The model training vectors are input into the initial model, and the initial model is trained using the mean squared error loss function to obtain the prediction model.

[0014] In one possible implementation, the step of collecting line information according to a predetermined period includes:

[0015] Receive line information reported by the metropolitan area network controller;

[0016] For each predetermined period, the line information received within a plurality of line information is averaged to obtain the line information for that predetermined period.

[0017] In one possible implementation, the step of inputting the current internet leased line information into a time-series-based prediction model to obtain the bandwidth utilization rate of the internet leased line at the next moment, as output by the prediction model, includes:

[0018] The time information in the current Internet leased line's line information is one-hot encoded and concatenated with the bandwidth utilization and network quality parameters at the current moment to obtain the model input vector;

[0019] The model input vector is input into the feature extraction layer of the prediction model to obtain the output feature vector corresponding to the current time.

[0020] The feature vector is input into the fully connected layer of the prediction model for regression prediction, and the bandwidth utilization of the next time step is obtained.

[0021] In one possible implementation, adjusting the internet leased line speed based on the bandwidth utilization rate of the internet leased line at the next time moment and the bandwidth utilization threshold of the internet leased line at the next time moment includes:

[0022] If the bandwidth utilization rate of the Internet leased line at the next moment is higher than the upper limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth boosting will be performed;

[0023] If the bandwidth utilization rate of the internet leased line at the next moment is lower than the lower limit of the bandwidth utilization rate of the internet leased line at the next moment, then bandwidth reduction will be implemented;

[0024] If the bandwidth utilization rate of the internet leased line at the next moment is between the upper and lower limits of the bandwidth utilization rate threshold for the internet leased line at the next moment, then the current bandwidth will be maintained.

[0025] On the other hand, this application provides an Internet leased line speed regulation device, including: an acquisition module, used to acquire the current Internet leased line line information, the line information including the current time information, bandwidth utilization rate and network quality parameters;

[0026] The prediction module is used to input the current Internet leased line information into a time-series-based prediction model to obtain the bandwidth utilization rate of the Internet leased line at the next moment, as output by the prediction model; wherein, the prediction model is a pre-trained model;

[0027] The speed control module is used to adjust the speed of the dedicated internet line based on the bandwidth utilization rate of the dedicated internet line at the next time moment and the bandwidth utilization rate threshold of the dedicated internet line at the next time moment.

[0028] In one possible implementation, the device further includes:

[0029] The training module is used to collect line information at a predetermined cycle within a predetermined data collection time, as feature parameters for model training. The feature parameters include time information, bandwidth utilization, and network quality parameters.

[0030] The time information in the feature parameters is one-hot encoded and concatenated with the corresponding feature parameters to obtain the model training vector;

[0031] An initial model is established, which is a time-series-based recurrent neural network model, including a feature extraction layer and a fully connected layer;

[0032] The model training vectors are input into the initial model, and the initial model is trained using the mean squared error loss function to obtain the prediction model.

[0033] In one possible implementation, the training module is specifically used for:

[0034] Receive line information reported by the metropolitan area network controller;

[0035] For each predetermined period, the line information received within a plurality of line information is averaged to obtain the line information for that predetermined period.

[0036] In one possible implementation, the speed control module is specifically used for:

[0037] The time information in the current Internet leased line's line information is one-hot encoded and concatenated with the bandwidth utilization and network quality parameters at the current moment to obtain the model input vector;

[0038] The model input vector is input into the feature extraction layer of the prediction model to obtain the output feature vector corresponding to the current time.

[0039] The feature vector is input into the fully connected layer of the prediction model for regression prediction, and the bandwidth utilization of the next time step is obtained.

[0040] In one possible implementation, the speed control module is specifically used for:

[0041] If the bandwidth utilization rate of the Internet leased line at the next moment is higher than the upper limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth boosting will be performed;

[0042] If the bandwidth utilization rate of the internet leased line at the next moment is lower than the lower limit of the bandwidth utilization rate of the internet leased line at the next moment, then bandwidth reduction will be implemented;

[0043] If the bandwidth utilization rate of the internet leased line at the next moment is between the upper and lower limits of the bandwidth utilization rate threshold for the internet leased line at the next moment, then the current bandwidth will be maintained.

[0044] In another aspect, this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the method described above.

[0045] In another aspect, this application provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, are used to implement the method described above.

[0046] The internet leased line speed adjustment method, apparatus, electronic device, and medium provided in this application use a regression prediction model based on a recurrent neural network to predict the bandwidth utilization rate of the user's internet leased line at the next moment. Based on the predicted bandwidth utilization rate of the internet leased line at the next moment and the corresponding bandwidth utilization threshold, the speed of the internet leased line is adjusted to meet the user's business needs for increased speed during busy periods and decreased speed during idle periods. The proposed solution uses an AI model to predict the traffic of the user's line and performs intelligent speed adjustment of the internet leased line, thereby achieving dynamic speed adjustment according to the customer's business needs. This avoids deterioration of line quality during busy periods and idle lines during off-peak periods, fully utilizing bandwidth resources and effectively improving user experience. Attached Figure Description

[0047] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0048] Figure 1 The diagram above illustrates a flowchart of the Internet leased line speed regulation method provided in Embodiment 1 of this application.

[0049] Figure 2 The diagram above illustrates a flowchart of the process for obtaining a prediction model provided in Embodiment 1 of this application.

[0050] Figure 3 The diagram below exemplarily illustrates the structure of the prediction model provided in Embodiment 1 of this application;

[0051] Figure 4 The diagram above illustrates a user's process for adjusting internet speed using a dedicated line, as provided in Embodiment 1 of this application.

[0052] Figure 5 The diagram above exemplarily illustrates the structure of the Internet leased line speed regulation device provided in Embodiment 2 of this application;

[0053] Figure 6 The diagram above illustrates the structure of the Internet leased line speed control electronic device provided in Embodiment 3 of this application.

[0054] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0055] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0056] In this application, a module refers to a functional module or a logical module. It can be in software form, where its function is implemented by a processor executing program code; or it can be in hardware form. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "OR" relationship.

[0057] In recent years, with the emergence of new technologies such as 5G, traditional home broadband services have become saturated, leading to new changes in the services carried by traditional metropolitan area networks. Internet leased line services have gradually emerged, creating new demands. Internet leased line services refer to services that connect customers' networks and devices to the Internet via dedicated links, providing customers with real-time online internet access at various speeds.

[0058] Traditional internet leased lines provide users with dedicated lines with fixed bandwidth, but during peak business periods, the line quality often deteriorates, affecting the user experience; while on holidays or at specific times, the lines are idle, increasing the cost of subscribing to leased lines.

[0059] Therefore, adjusting dedicated line speeds based on customer-side traffic usage can better meet customer needs. Furthermore, with the development of artificial intelligence technology, AI-based user behavior prediction is increasingly integrated into people's daily lives. For enterprise users of dedicated internet lines, user traffic usage within specific time periods often exhibits certain patterns, providing a basis for training user traffic prediction models.

[0060] The technical solutions of this application are illustrated below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0061] Example 1

[0062] Figure 1 This is a flowchart illustrating an embodiment of the internet leased line speed adjustment method provided in this application. Figure 1 As shown, the internet leased line speed adjustment method provided in this embodiment may include:

[0063] S101, Obtain the current Internet leased line information, the line information including the current time information, bandwidth utilization and network quality parameters;

[0064] S102, the line information of the Internet leased line is input into the time-series-based prediction model to obtain the bandwidth utilization rate of the Internet leased line at the next moment output by the prediction model; wherein, the prediction model is a pre-trained model;

[0065] S103, based on the bandwidth utilization rate of the Internet leased line at the next time moment and the bandwidth utilization rate threshold of the Internet leased line at the next time moment, the leased line speed is adjusted.

[0066] In practical applications, the execution subject of this embodiment can be an Internet leased line speed regulation device, which can be implemented by a computer program, such as application software; or it can be implemented as a medium storing relevant computer programs, such as a USB flash drive or cloud drive; or it can be implemented by a physical device that integrates or installs relevant computer programs, such as a chip or server.

[0067] Surveys of dedicated internet line users reveal a pattern in the busy and idle periods of their internet lines. For example, restaurants typically experience peak traffic and higher bandwidth utilization, resulting in lower network quality, between 11:00 AM and 2:00 PM and between 5:00 PM and 8:00 PM. During the remaining hours, the lines are less busy and the network quality is better. Therefore, the busy and idle periods for specific businesses exhibit regularity, and this pattern can be learned through deep learning models to predict line traffic. In this context, this embodiment introduces an AI model to predict user line traffic, thereby determining peak and off-peak times, and then intelligently adjusts the network speed based on the results.

[0068] Specifically, the internet leased line speed control device acquires the current internet leased line information, which includes the current time, bandwidth utilization, and network quality parameters. For example, the line information may include time, day of the week, bandwidth utilization, and Quality of Service (QoS) (which may include latency, jitter, packet loss, etc.), and can be selected according to the actual application situation; no restrictions are imposed here. The obtained current internet leased line information is input into the prediction model to obtain the bandwidth utilization of the internet leased line at the next time moment. The prediction model is a time-series recurrent neural network regression prediction model, which has been pre-trained.

[0069] For example, suppose a catering company orders the dedicated internet line speed adjustment product offered in this solution. The dedicated internet line speed adjustment device automatically collects the company's dedicated internet line information, inputs the current line information into a trained prediction model, and obtains the predicted line information for the next moment. For instance, if the current line information is Saturday at 10:00 AM with a bandwidth utilization of 70%, inputting it into the prediction model will yield the next line information as Saturday at 11:00 AM with a bandwidth utilization of 100%. The preset bandwidth utilization threshold is 75%, indicating that the line will enter a busy state in the next moment, requiring a dedicated line speed increase.

[0070] In practical applications, there are multiple ways to obtain a prediction model. Figure 2 The diagram below illustrates a process for obtaining a prediction model according to an embodiment of this application. In one example, the method further includes:

[0071] S201, During the predetermined data collection time, collect line information according to a predetermined cycle as feature parameters for model training. The feature parameters include time information, bandwidth utilization, and network quality parameters.

[0072] S202, perform one-hot encoding on the time information in the feature parameters and concatenate it with the corresponding feature parameters to obtain the model training vector;

[0073] S203, Establish an initial model, which is a time-series-based recurrent neural network model, including a feature extraction layer and a fully connected layer;

[0074] S204, input the model training vector into the initial model, and use the mean squared error loss function to train the initial model to obtain the prediction model.

[0075] Specifically, this solution uses a time-series-based recurrent neural network model as the initial model. This initial model includes a feature extraction layer for feature extraction and a fully connected layer for regression prediction. The user presets the data collection time and period. The dedicated internet line speed control device collects line information according to the preset period and period to obtain historical line information, which is used as feature parameters for model training. These feature parameters include, but are not limited to, time information (time, day of the week), bandwidth utilization, and network quality parameters (latency, jitter, packet loss), which can be selected according to the needs of the actual application. Since the time information (time, day of the week) is discrete and the data volume is moderate, it is uniquely encoded to extend the values ​​of the discrete features to Euclidean space, enabling better feature analysis. The processed time information is concatenated with other feature parameters to obtain the model training vector, which is used by the model to learn the user's traffic usage patterns. The model training vector is input into the initial model. To improve the accuracy of regression prediction, a mean squared error loss function is used to train the initial model, resulting in the prediction model.

[0076] In one example, the step of inputting the current internet leased line information into a time-series-based prediction model to obtain the bandwidth utilization rate of the internet leased line at the next moment, as output by the prediction model, includes:

[0077] The time information in the current Internet leased line's line information is one-hot encoded and concatenated with the bandwidth utilization and network quality parameters at the current moment to obtain the model input vector;

[0078] The model input vector is input into the feature extraction layer of the prediction model to obtain the output feature vector corresponding to the current time.

[0079] The feature vector is input into the fully connected layer of the prediction model for regression prediction, and the bandwidth utilization of the next time step is obtained.

[0080] Specifically, based on the current line information, the model input vector is obtained, and the process is similar to that of obtaining the model training vector, so it will not be repeated here. After the model input vector is input into the prediction model, the feature extraction layer extracts features from the model input vector to obtain the feature vector corresponding to the current time step. The feature vector corresponding to the current time step contains the line feature information at the current time step and historical feature information. The feature vector corresponding to the current time step is input into the fully connected layer, and the bandwidth utilization predicted for the next time step is obtained through regression prediction.

[0081] Figure 3 This is a structural example diagram of a prediction model provided in an embodiment of this application. Figure 3 As shown, assuming user line usage is collected every hour, line information is obtained, including: time (0-23), day of the week (1-7), user bandwidth utilization, and user QoS (latency, jitter, packet loss). Feature processing is performed on the line information at different times: one-hot encoding is used to represent the time and day of the week parameters, while the user bandwidth utilization and QoS values ​​are data values ​​and therefore do not require mapping. The processed line information from different times is concatenated to obtain the vector X corresponding to each time moment. i .

[0082] X i =Concat[Onehot(F i ):Onehot(F week ):F Bi-1 :F QoSi-1 ]

[0083] Let time t represent the next time step that needs to be predicted, and time t-1 represent the current time corresponding to the next time step. If i = t, then the model input vector X... t The calculation formula is as follows:

[0084] X t =Concat[Onehot(F t ):Onehot(F week ):F Bt-1 :F QoSt-1 ]

[0085] Similarly, based on historical route information, the model training vectors X0, X1, ..., X can be obtained. t-1 .

[0086] Since the goal is to predict user bandwidth utilization at the next time step, a temporal recurrent neural network (RNN) model is chosen for modeling, with gated recurrent units (GRUs) selected as the RRN units. Because the final task is regression prediction, the mean squared error (MSE) loss function is used for model training, calculated as follows:

[0087] Loss = MSE(o t ,F Bt )=(o i -F Bi ) 2 / n,,i=1,2,3…n

[0088] Through the model input vector X t and the hidden state vector h corresponding to the current time step obtained during training. t-1 To obtain the feature vector h that contains current input information and historical feature information. t Predict the eigenvector h corresponding to the next time step t. t The calculation formula is as follows:

[0089] h t =GRU(h t-1 ,X t )

[0090] Obtain the eigenvector h t Then, a fully connected neural network is used to perform regression prediction output to predict the bandwidth utilization rate at the next time step t. t The calculation formula is:

[0091] o t =MLP(h t )

[0092] MLP stands for Multilayer Fully Connected Neural Network.

[0093] Ultimately, the AI ​​model designed using this method was able to predict the user bandwidth utilization for the next moment based on the user's historical line traffic information and the network status information at the current moment.

[0094] In one example, the collection of line information according to a predetermined period includes:

[0095] Receive line information reported by the metropolitan area network controller;

[0096] For each predetermined period, the line information received within a plurality of line information is averaged to obtain the line information for that predetermined period.

[0097] Traditional internet leased lines are carried by municipal IP metropolitan area networks (MANs) and IP RAN networks, with each municipality having its own MAN and RAN. To achieve speed adjustment for customers nationwide, it's necessary to connect to the network management systems of each province separately, increasing the complexity of the task. To meet the intelligent speed adjustment needs of internet leased line users, this solution uses an intelligent metropolitan area network (MAN) to carry the internet leased lines. An intelligent metropolitan area network is a converged network of municipal IP MANs and IP RANs, merging the networks of all cities into a single intelligent metropolitan area network. To achieve speed adjustment, only the intelligent metropolitan area network SDN controller needs to be connected, thus reducing the complexity of intelligent speed adjustment task integration. The intelligent metropolitan area network is a fully SDN-enabled network; all device configurations can be distributed through the intelligent metropolitan area network SDN controller, and configuration distribution can be completed within minutes. Therefore, internet leased lines within the intelligent metropolitan area network are capable of distributing speed adjustment tasks seamlessly to users.

[0098] Specifically, the collected line information is reported to the intelligent metropolitan area network (MAN) SDN controller for storage and processing. The MAN SDN controller then reports the processed user line information to the internet leased line speed control device. Within each predetermined period, the internet leased line speed control device receives line information multiple times, and the average value is taken as the line information for that period. For example, if the internet leased line speed control device collects the status of user lines every hour, and the minimum granularity of the MAN controller's line status reporting is 15 minutes, then the internet leased line speed control device takes the average of four received controller reports within each period as the line information for each hour.

[0099] In practical applications, there can be various methods for adjusting the speed of a leased internet line. In one example, adjusting the speed of the leased internet line based on its bandwidth utilization rate at the next moment and the bandwidth utilization rate threshold at the next moment includes:

[0100] If the bandwidth utilization rate of the Internet leased line at the next moment is higher than the upper limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth speed-up will be performed;

[0101] If the bandwidth utilization rate of the Internet leased line at the next moment is lower than the lower limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth reduction will be implemented;

[0102] If the bandwidth utilization rate of the internet leased line at the next moment is between the upper and lower limits of the bandwidth utilization rate threshold for the internet leased line at the next moment, then the current bandwidth will be maintained.

[0103] Specifically, users can pre-set speed adjustment rules, including: upper and lower limits of bandwidth utilization thresholds for speed increases and decreases, and the bandwidth for speed increases and decreases when the speed adjustment thresholds are triggered. Based on the predicted bandwidth utilization of the dedicated internet line in the next moment, the relationship between this bandwidth utilization and the upper and lower thresholds of the dedicated internet line in the next moment is determined. If the bandwidth utilization is higher than the preset upper threshold, it indicates that the next moment is a busy period, and the bandwidth is increased to the bandwidth set in the speed adjustment rules; if the bandwidth utilization is lower than the preset lower threshold, it indicates that the next moment is an idle period, and the line will be idle, so the bandwidth is decreased to the bandwidth set in the speed adjustment rules; if the bandwidth utilization is between the preset upper and lower thresholds, it indicates that the line usage is good in the next moment, and the original bandwidth remains unchanged.

[0104] For example, suppose a user sets a bandwidth utilization threshold of 70% upper limit and 30% lower limit. After speed increase, the bandwidth is 150% of the original bandwidth, and after speed decrease, the bandwidth is 50% of the original bandwidth. If it is predicted that the user's line bandwidth utilization will be 90% in the next moment, the bandwidth will be adjusted to 150% of the original bandwidth; if it is predicted that the user's line bandwidth utilization will be 20% in the next moment, the bandwidth will be adjusted to 50% of the original bandwidth; if it is predicted that the user's line bandwidth utilization will be 50% in the next moment, the original bandwidth will remain unchanged.

[0105] Figure 4 This is a schematic diagram illustrating a user's process of adjusting internet speed using a dedicated line, as provided in one embodiment of this application. Figure 4 As shown, users log into the self-service system, select their dedicated internet line, and choose intelligent speed control. They then set the data collection period (minimum 15 days) and speed control rules (upper and lower limits of bandwidth utilization thresholds for speed increases and decreases, and the bandwidth for speed increases and decreases when the speed control threshold is triggered). The intelligent training mode is activated, and the dedicated internet line speed control device acquires the user's dedicated line status information reported by the controller within a preset time period, and sends the obtained feature information to the AI ​​model for training. The latest training model is loaded, and the user's historical dedicated internet line status information for the current day is collected and sent to the AI ​​model to predict the user's bandwidth utilization at the next moment. If the predicted bandwidth utilization is higher than the user-designed upper threshold, the system automatically issues an increase configuration, with the increase bandwidth set in the user's rules; if the predicted bandwidth utilization is lower than the user-designed lower threshold, the system automatically issues a decrease configuration, with the decrease bandwidth also preset by the user; if the predicted bandwidth is between the user-set upper and lower thresholds, the original bandwidth remains unchanged.

[0106] The internet leased line speed adjustment method provided in this embodiment uses a regression prediction model based on a recurrent neural network to predict the bandwidth utilization rate of the user's internet leased line at the next moment. Based on the predicted bandwidth utilization rate of the internet leased line at the next moment and the corresponding bandwidth utilization threshold, the speed of the internet leased line is adjusted to meet the user's needs for increased speed during busy periods and decreased speed during idle periods. The solution proposed in this application uses an AI model to predict the traffic of the user's line and performs intelligent speed adjustment of the internet leased line, thereby achieving dynamic speed adjustment according to the customer's business needs. This avoids deterioration of line quality during busy periods and idle lines during off-peak periods, fully utilizing bandwidth resources and effectively improving user experience.

[0107] Example 2

[0108] Figure 5 This is a schematic diagram of the structure of an internet leased line speed regulation device provided in one embodiment of this application. Figure 5 As shown, the internet leased line speed regulation device provided in this embodiment may include:

[0109] The acquisition module 51 is used to acquire the current Internet leased line information, which includes the current time information, bandwidth utilization, and network quality parameters.

[0110] The prediction module 52 is used to input the line information of the Internet leased line into the time-series-based prediction model to obtain the bandwidth utilization rate of the Internet leased line at the next moment output by the prediction model; wherein, the prediction model is a pre-trained model.

[0111] Speed ​​control module 53 is used to adjust the speed of the Internet leased line according to the bandwidth utilization rate of the Internet leased line at the next time moment and the bandwidth utilization rate threshold of the Internet leased line at the next time moment.

[0112] In practical applications, internet leased line speed control devices can be implemented through computer programs, such as application software; or they can be implemented as media storing relevant computer programs, such as USB flash drives or cloud drives; or they can be implemented through physical devices that integrate or install relevant computer programs, such as chips or servers.

[0113] Surveys of dedicated internet line users reveal a pattern in the busy and idle periods of their internet lines. For example, restaurants typically experience peak traffic and higher bandwidth utilization, resulting in lower network quality, between 11:00 AM and 2:00 PM and between 5:00 PM and 8:00 PM. During the remaining hours, the lines are less busy and the network quality is better. Therefore, the busy and idle periods for specific businesses exhibit regularity, and this pattern can be learned through deep learning models to predict line traffic. In this context, this embodiment introduces an AI model to predict user line traffic, thereby determining peak and off-peak times, and then intelligently adjusts the network speed based on the results.

[0114] Specifically, the internet leased line speed control device acquires the current internet leased line information, which includes the current time, bandwidth utilization, and network quality parameters. For example, the line information may include time, day of the week, bandwidth utilization, and Quality of Service (QoS) (which may include latency, jitter, packet loss, etc.), and can be selected according to the actual application situation; no restrictions are imposed here. The obtained current internet leased line information is input into the prediction model to obtain the bandwidth utilization of the internet leased line at the next time moment. The prediction model is a time-series recurrent neural network regression prediction model, which has been pre-trained.

[0115] In practical applications, there can be multiple choices for the prediction model. In one example, the device further includes:

[0116] The training module is used to collect line information at a predetermined cycle within a predetermined data collection time, as feature parameters for model training. The feature parameters include time information, bandwidth utilization, and network quality parameters.

[0117] The time information in the feature parameters is one-hot encoded and concatenated with the corresponding feature parameters to obtain the model training vector;

[0118] An initial model is established, which is a time-series-based recurrent neural network model, including a feature extraction layer and a fully connected layer;

[0119] The model training vectors are input into the initial model, and the initial model is trained using the mean squared error loss function to obtain the prediction model.

[0120] Specifically, this solution uses a time-series-based recurrent neural network model as the initial model. This initial model includes a feature extraction layer for feature extraction and a fully connected layer for regression prediction. The user presets the data collection time and period. The dedicated internet line speed control device collects line information according to the preset period and period to obtain historical line information, which is used as feature parameters for model training. These feature parameters include, but are not limited to, time information (time, day of the week), bandwidth utilization, and network quality parameters (latency, jitter, packet loss), which can be selected according to the needs of the actual application. Since the time information (time, day of the week) is discrete and the data volume is moderate, it is uniquely encoded to extend the values ​​of the discrete features to Euclidean space, enabling better feature analysis. The processed time information is concatenated with other feature parameters to obtain the model training vector, which is used by the model to learn the user's traffic usage patterns. The model training vector is input into the initial model. To improve the accuracy of regression prediction, a mean squared error loss function is used to train the initial model, resulting in the prediction model.

[0121] In one example, the prediction module 52 is specifically used for:

[0122] The time information in the current Internet leased line's line information is one-hot encoded and concatenated with the bandwidth utilization and network quality parameters at the current moment to obtain the model input vector;

[0123] The model input vector is input into the feature extraction layer of the prediction model to obtain the output feature vector corresponding to the current time.

[0124] The feature vector is input into the fully connected layer of the prediction model for regression prediction, and the bandwidth utilization of the next time step is obtained.

[0125] Specifically, based on the current line information, the model input vector is obtained, and the process is similar to that of obtaining the model training vector, so it will not be repeated here. After the model input vector is input into the prediction model, the feature extraction layer extracts features from the model input vector to obtain the feature vector corresponding to the current time step. The feature vector corresponding to the current time step contains the line feature information at the current time step and historical feature information. The feature vector corresponding to the current time step is input into the fully connected layer, and the bandwidth utilization predicted for the next time step is obtained through regression prediction.

[0126] In one example, the training module is specifically used for:

[0127] Receive line information reported by the metropolitan area network controller;

[0128] For each predetermined period, the line information received within a plurality of line information is averaged to obtain the line information for that predetermined period.

[0129] Traditional internet leased lines are carried by municipal IP metropolitan area networks (MANs) and IP RAN networks, with each municipality having its own MAN and RAN. To achieve speed adjustment for customers nationwide, it's necessary to connect to the network management systems of each province separately, increasing the complexity of the task. To meet the intelligent speed adjustment needs of internet leased line users, this solution uses an intelligent metropolitan area network (MAN) to carry the internet leased lines. An intelligent metropolitan area network is a converged network of municipal IP MANs and IP RANs, merging the networks of all cities into a single intelligent metropolitan area network. To achieve speed adjustment, only the intelligent metropolitan area network SDN controller needs to be connected, thus reducing the complexity of intelligent speed adjustment task integration. The intelligent metropolitan area network is a fully SDN-enabled network; all device configurations can be distributed through the intelligent metropolitan area network SDN controller, and configuration distribution can be completed within minutes. Therefore, internet leased lines within the intelligent metropolitan area network are capable of distributing speed adjustment tasks seamlessly to users.

[0130] Specifically, the collected line information will be reported to the intelligent metropolitan area network SDN controller for storage and processing. The intelligent metropolitan area network SDN controller will then report the processed user line information to the internet leased line speed control device. Within each predetermined period, the internet leased line speed control device will receive line information multiple times, and the average value will be taken as the line information for that period.

[0131] In practical applications, there can be various methods for speed regulation on dedicated lines. In one example, the speed regulation module 53 is specifically used for:

[0132] If the bandwidth utilization rate of the Internet leased line at the next moment is higher than the upper limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth boosting will be performed;

[0133] If the bandwidth utilization rate of the internet leased line at the next moment is lower than the lower limit of the bandwidth utilization rate of the internet leased line at the next moment, then bandwidth reduction will be implemented;

[0134] If the bandwidth utilization rate of the internet leased line at the next moment is between the upper and lower limits of the bandwidth utilization rate threshold for the internet leased line at the next moment, then the current bandwidth will be maintained.

[0135] Specifically, users can pre-set speed adjustment rules, including: upper and lower limits of bandwidth utilization thresholds for speed increases and decreases, and the bandwidth for speed increases and decreases when the speed adjustment thresholds are triggered. Based on the predicted bandwidth utilization of the dedicated internet line in the next moment, the relationship between this bandwidth utilization and the upper and lower thresholds of the dedicated internet line in the next moment is determined. If the bandwidth utilization is higher than the preset upper threshold, it indicates that the next moment is a busy period, and the bandwidth is increased to the bandwidth set in the speed adjustment rules; if the bandwidth utilization is lower than the preset lower threshold, it indicates that the next moment is an idle period, and the line will be in an idle state, and the bandwidth is decreased to the bandwidth set in the speed adjustment rules; if the bandwidth utilization is between the preset upper and lower thresholds, it indicates that the line usage is good in the next moment, and the original bandwidth remains unchanged.

[0136] The internet leased line speed regulation device provided in this embodiment uses a regression prediction model based on a recurrent neural network to predict the bandwidth utilization rate of the user's internet leased line at the next moment. Based on the predicted bandwidth utilization rate of the internet leased line at the next moment and the corresponding bandwidth utilization threshold, the device regulates the speed of the internet leased line to meet the user's needs for increased speed during busy periods and decreased speed during idle periods. The solution proposed in this application uses an AI model to predict the traffic of the user's line and performs intelligent speed regulation of the internet leased line. This achieves dynamic speed regulation based on customer-side business needs, avoiding deterioration of line quality during busy periods and line idleness during off-peak periods, fully utilizing bandwidth resources and effectively improving user experience.

[0137] Example 3

[0138] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure, such as... Figure 6 As shown, the electronic device includes:

[0139] The electronic device includes a processor 291 and a memory 292; it may also include a communication interface 293 and a bus 294. The processor 291, memory 292, and communication interface 293 can communicate with each other via the bus 294. The communication interface 293 can be used for information transmission. The processor 291 can invoke logical instructions stored in the memory 292 to execute the methods of the above embodiments.

[0140] Furthermore, the logic instructions in the aforementioned memory 292 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0141] The memory 292, as a computer-readable storage medium, can be used to store software programs and computer-executable programs, such as program instructions / modules corresponding to the methods in the embodiments of this disclosure. The processor 291 executes functional applications and data processing by running the software programs, instructions, and modules stored in the memory 292, thereby implementing the methods in the above-described method embodiments.

[0142] The memory 292 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory 292 may include high-speed random access memory and may also include non-volatile memory.

[0143] This disclosure provides a non-transitory computer-readable storage medium storing computer-executable instructions that, when executed by a processor, are used to implement the methods described in the foregoing embodiments.

[0144] Example 4

[0145] This disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the methods provided in any of the embodiments described above.

[0146] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0147] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for adjusting the speed of a dedicated internet line, characterized in that, include: During the predetermined data collection period, historical line information is collected according to a predetermined cycle. For each predetermined cycle, the historical line information for that cycle is obtained by averaging multiple reported information received from the metropolitan area network controller. The historical line information includes time information, bandwidth utilization, and network quality parameters for the corresponding cycle. The time information in the historical line information is one-hot encoded and concatenated with the corresponding bandwidth utilization and network quality parameters to obtain the model training vector; an initial model is established, which is a time-series based recurrent neural network model, including a feature extraction layer and a fully connected layer; the model training vector is input into the initial model, and the model is trained using the mean squared error loss function to obtain the prediction model; Obtain the current line information of the current Internet leased line, which includes the current time information, the current bandwidth utilization rate, and the current network quality parameters; The time information in the current line information is one-hot encoded and concatenated with the bandwidth utilization and network quality parameters at the current moment to obtain the model input vector; the model input vector is input into the feature extraction layer of the prediction model to obtain the output feature vector corresponding to the current moment; the feature vector is input into the fully connected layer of the prediction model for regression prediction to obtain the output bandwidth utilization at the next moment. Based on the bandwidth utilization rate of the internet leased line at the next moment and the pre-configured bandwidth utilization rate threshold corresponding to the time information of the next moment, the internet leased line speed is adjusted.

2. The method according to claim 1, characterized in that, The step of adjusting the dedicated internet line speed based on the bandwidth utilization rate of the dedicated internet line at the next moment and a pre-configured bandwidth utilization rate threshold corresponding to the time information of the next moment includes: If the bandwidth utilization rate of the Internet leased line at the next moment is higher than the upper limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth boosting will be performed; If the bandwidth utilization rate of the Internet leased line at the next moment is lower than the lower limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth reduction will be implemented; If the bandwidth utilization rate of the internet leased line at the next moment is between the upper and lower limits of the bandwidth utilization rate threshold for the internet leased line at the next moment, then the current bandwidth will be maintained.

3. A speed control device for internet leased lines, characterized in that, include: The acquisition module is used to acquire the current line information of the current Internet leased line. The current line information includes the current time information, the current bandwidth utilization rate, and the current network quality parameters. The prediction module is used to perform one-hot encoding on the time information in the current line information, and concatenate it with the bandwidth utilization and network quality parameters at the current moment to obtain a model input vector; the model input vector is input into the feature extraction layer of the prediction model to obtain the output feature vector corresponding to the current moment; the feature vector is input into the fully connected layer of the prediction model for regression prediction to obtain the output bandwidth utilization at the next moment; wherein, the prediction model is a model pre-trained based on the following method: collecting historical line information according to a predetermined period within a predetermined data collection time; wherein, for each predetermined period, by analyzing the received data... The historical line information for the predetermined period is obtained by averaging multiple reported information from the metropolitan area network controller. This historical line information includes time information, bandwidth utilization, and network quality parameters for the corresponding period. The time information in the historical line information is one-hot encoded and concatenated with the corresponding bandwidth utilization and network quality parameters to obtain a model training vector. An initial model is established, which is a time-series-based recurrent neural network model, including a feature extraction layer and a fully connected layer. The model training vector is input into the initial model, and the model is trained using a mean squared error loss function to obtain the prediction model. The speed control module is used to adjust the speed of the internet leased line based on the bandwidth utilization rate of the internet leased line at the next moment and a pre-configured bandwidth utilization rate threshold corresponding to the time information of the next moment.

4. The apparatus according to claim 3, characterized in that, The speed control module is specifically used for: If the bandwidth utilization rate of the Internet leased line at the next moment is higher than the upper limit of the bandwidth utilization rate of the Internet leased line at the next moment, then bandwidth boosting will be performed; If the bandwidth utilization rate of the internet leased line at the next moment is lower than the lower limit of the bandwidth utilization rate of the internet leased line at the next moment, then bandwidth reduction will be implemented; If the bandwidth utilization rate of the internet leased line at the next moment is between the upper and lower limits of the bandwidth utilization rate threshold for the internet leased line at the next moment, then the current bandwidth will be maintained.

5. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in claim 1 or 2.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in claim 1 or 2.

Images