Explicit congestion notification dynamic optimization method and device based on long short-term memory network, equipment and storage medium

Abstract

The invention discloses an explicit congestion notification dynamic optimization method, device and equipment based on a long short-term memory network and a storage medium, and relates to the technical field of network optimization, and the method comprises the steps: obtaining real-time network flow data; generating a dynamic congestion threshold value through the long-short-term memory network and the real-time network flow data; and displaying a congestion notification mark based on the dynamic congestion threshold. The invention provides a method for dynamically optimizing explicit congestion notification parameters through combination of a long-short term memory network and explicit congestion notification. According to the invention, the long-term dependency relationship of real-time network flow data is captured through the long-short-term memory network, high-precision prediction of the network congestion condition is realized, and the network congestion threshold is dynamically adjusted, so that the display congestion notification marking based on the dynamic congestion threshold effectively adapts to a complex and changeable network environment, and the network congestion notification marking efficiency is improved. And the dynamic self-adaptive capability is improved.

Description

Technical Field

[0001] This application relates to the field of network optimization technologies, and in particular, to an explicit congestion notification dynamic optimization method, apparatus, device, and storage medium based on a long short-term memory network. Background Art

[0002] With the explosive growth of Internet traffic, network congestion management has become a key issue in improving network service quality (QoS). Explicit Congestion Notification (ECN) is a mechanism for notifying network congestion by marking data packets, but its traditional implementation relies on static thresholds (such as fixed queue lengths) to trigger congestion marking and cannot adapt to dynamic network environments.

[0003] Therefore, the static threshold mechanism of traditional ECN has problems such as response lag and poor flexibility in dynamic network environments. How to provide an explicit congestion notification method with strong adaptability has become an urgent problem to be solved. Summary of the Invention

[0004] The main purpose of this application is to provide an explicit congestion notification dynamic optimization method, apparatus, device, and storage medium based on a long short-term memory network, aiming to solve the technical problem of how to provide an explicit congestion notification method with strong adaptability.

[0005] To achieve the above object, this application proposes an explicit congestion notification dynamic optimization method based on a long short-term memory network, and the method includes:

[0006] Obtain real-time network traffic data;

[0007] Generate a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data;

[0008] Perform explicit congestion notification marking based on the dynamic congestion threshold.

[0009] In one embodiment, the step of generating a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data includes:

[0010] Perform data preprocessing on the real-time traffic data to generate a time series data set;

[0011] Generate a dynamic congestion threshold through a long short-term memory network and the time series data set.

[0012] In one embodiment, the step of generating a dynamic congestion threshold through a long short-term memory network and the time series data set includes:

[0013] Input the time series data set into a long short - term memory network for predicting the network congestion probability to obtain the dynamic congestion probability;

[0014] Determine the dynamic congestion threshold according to the dynamic congestion probability.

[0015] In one embodiment, the step of determining the dynamic congestion threshold according to the dynamic congestion probability includes:

[0016] Obtain the standard congestion threshold output by the long short - term memory network;

[0017] Calculate the dynamic congestion threshold according to the standard congestion threshold and the dynamic congestion probability;

[0018] Threshold=Base_Threshold×(1 - α×Predicted_Congestion_Probability);

[0019] In the formula, Threshold is the dynamic congestion threshold, Base_Threshold is the standard congestion threshold, Predicted_Congestion_Probability is the dynamic congestion probability, and α is the adjustment coefficient.

[0020] In one embodiment, the step of pre - processing the real - time traffic data to generate a time series data set includes:

[0021] Normalize the real - time traffic data to obtain the standard traffic data;

[0022] Divide the standard traffic data according to a preset window length to generate a time series data set.

[0023] In one embodiment, after displaying the congestion notification mark based on the dynamic congestion threshold, it further includes:

[0024] Obtain the marked network performance data;

[0025] Iteratively optimize the long short - term memory network based on the network performance data and a preset loss function.

[0026] In one embodiment, the step of iteratively optimizing the long short - term memory network based on the network performance data and a preset loss function includes:

[0027] Determine the current adjustment error based on the network performance data and the expected performance data;

[0028] If it is detected that the deviation duration between the current adjustment error and the preset adjustment error exceeds the preset deviation duration, the long short-term memory network is iteratively optimized through a preset loss function.

[0029] In addition, to achieve the above object, the present application also proposes an explicit congestion notification dynamic optimization device based on a long short-term memory network. The explicit congestion notification dynamic optimization device based on a long short-term memory network includes:

[0030] A data acquisition module, configured to acquire real-time network traffic data;

[0031] A data analysis module, configured to generate a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data;

[0032] An optimization module, configured to perform explicit congestion notification marking based on the dynamic congestion threshold.

[0033] In addition, to achieve the above object, the present application also proposes an explicit congestion notification dynamic optimization device based on a long short-term memory network. The device includes: a memory, a processor, and an explicit congestion notification dynamic optimization program stored on the memory and executable on the processor. The explicit congestion notification dynamic optimization program based on a long short-term memory network is configured to implement the steps of the explicit congestion notification dynamic optimization method based on a long short-term memory network as described above.

[0034] In addition, to achieve the above object, the present application also provides a storage medium. The storage medium is a computer-readable storage medium, and a program for implementing the explicit congestion notification dynamic optimization method based on a long short-term memory network is stored on the computer-readable storage medium. The program for implementing the explicit congestion notification dynamic optimization method based on a long short-term memory network is executed by a processor to implement the steps of the explicit congestion notification dynamic optimization method based on a long short-term memory network as described above.

[0035] The present application provides an explicit congestion notification dynamic optimization method, device, device, and storage medium based on a long short-term memory network. The method includes acquiring real-time network traffic data; generating a dynamic congestion threshold through a long short-term memory network and real-time network traffic data; and performing explicit congestion notification marking based on the dynamic congestion threshold. The present application proposes a method for dynamically optimizing explicit congestion notification parameters by combining a long short-term memory network and explicit congestion notification. The present application captures the long-term dependencies of real-time network traffic data through a long short-term memory network, realizes high-precision prediction of network congestion conditions, dynamically adjusts the network congestion threshold, so that the explicit congestion notification marking based on the dynamic congestion threshold can effectively adapt to complex and changeable network environments and improve its dynamic adaptive ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The accompanying drawings here are incorporated into the description and form a part of this description, showing embodiments consistent with the present application, and are used together with the description to explain the principles of the present application.

[0037] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0038] Figure 1 It is the first process schematic diagram of the first embodiment of the explicit congestion notification dynamic optimization method based on the long short-term memory network of the present application;

[0039] Figure 2 It is the second process schematic diagram of the first embodiment of the explicit congestion notification dynamic optimization method based on the long short-term memory network of the present application;

[0040] Figure 3 It is the process schematic diagram of the second embodiment of the explicit congestion notification dynamic optimization method based on the long short-term memory network of the present application;

[0041] Figure 4 It is the brief process schematic diagram of the explicit congestion notification dynamic optimization method based on the long short-term memory network of the present application;

[0042] Figure 5 It is the module structure schematic diagram of the explicit congestion notification dynamic optimization device based on the long short-term memory network in the embodiments of the present application;

[0043] Figure 6 It is the device structure schematic diagram of the hardware operating environment involved in the explicit congestion notification dynamic optimization method based on the long short-term memory network in the embodiments of the present application.

[0044] The realization of the purpose, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. Detailed Embodiments

[0045] It should be understood that the specific embodiments described here are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0046] To better understand the technical solutions of the present application, the following will be described in detail in combination with the accompanying drawings of the description and the specific embodiments.

[0047] The main solution of the present application is: obtaining real-time network traffic data; generating a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data; performing display congestion notification marking based on the dynamic congestion threshold.

[0048] With the explosive growth of Internet traffic, network congestion management has become a key issue in improving network quality of service (QoS). Explicit congestion notification (ECN) is a mechanism for notifying network congestion by marking data packets. However, its traditional implementation relies on static thresholds (such as fixed queue lengths) to trigger congestion marking and cannot adapt to the dynamically changing network environment. Therefore, the static threshold mechanism of traditional ECN has problems such as response lag and poor flexibility in a dynamic network environment. There is an urgent need for a technical solution that can predict network congestion in real time and dynamically adjust the threshold.

[0049] To solve the above problems, this application combines long short-term memory networks to analyze real-time network traffic data to generate dynamic congestion thresholds, and then generates explicit congestion notification marked data packets based on the dynamic congestion thresholds, so as to perform explicit congestion notification based on the optimized explicit congestion notification marked data packets, thereby realizing dynamic adjustment of the explicit congestion notification threshold according to the real-time network traffic, avoiding premature or late triggering of congestion notifications and affecting network efficiency, and also avoiding waste or over-occupation of network resources, thereby improving the adaptability of the solution.

[0050] It should be noted that the execution entity of this embodiment can be an explicit congestion notification dynamic optimization system based on long short-term memory networks, or a computing service device with data processing, network communication, and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an explicit congestion notification dynamic optimization device based on long short-term memory networks that can implement the above functions, etc. This embodiment does not make specific limitations in this regard. The following takes an explicit congestion notification dynamic optimization device based on long short-term memory networks (referred to as the optimization device for short) as the execution entity as an example to illustrate this embodiment and the following embodiments.

[0051] Based on this, an embodiment of this application provides an explicit congestion notification dynamic optimization method based on long short-term memory networks, referring to Figure 1 , Figure 1 is the first process schematic diagram of the first embodiment of the explicit congestion notification dynamic optimization method based on long short-term memory networks of this application.

[0052] In this embodiment, the explicit congestion notification dynamic optimization method based on long short-term memory networks includes steps S10 to S30:

[0053] Step S10, obtain real-time network traffic data;

[0054] It is easy to understand that the above real-time network traffic data can be real-time traffic data collected from a data center or network device (such as a switch) to be optimized, and the data includes but is not limited to indicators such as bandwidth utilization rate, queue length, throughput, latency, CPU load, memory occupancy rate, or TCP retransmission rate.

[0055] Step S20: Generate a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data;

[0056] Step S30: Based on the dynamic congestion threshold, perform display of congestion notification marks.

[0057] It should be understood that in the prior art, the static threshold setting of ECN (Explicit Congestion Notification) cannot be dynamically adjusted according to the real-time traffic pattern, which easily leads to premature or late triggering of congestion notifications and affects network efficiency. At the same time, the fixed threshold may cause waste or over-occupation of network resources (such as bandwidth, buffer), reducing the overall performance. Therefore, in this embodiment, to improve the adaptability of explicit congestion notification and based on the characteristic that the traffic in network communication is strongly correlated with time, compared with convolutional neural networks and other models, in this embodiment, the real-time network traffic data can be input into a long short-term memory network (LSTM), and then the LSTM model can capture the traffic time series dependence relationship through the hidden state to dynamically predict an appropriate dynamic congestion threshold.

[0058] In a feasible implementation manner, in this embodiment, step S20 may include steps A1 to A2:

[0059] Step A1: Perform data preprocessing on the real-time traffic data to generate a time series data set;

[0060] Step A2: Generate a dynamic congestion threshold through a long short-term memory network and the time series data set.

[0061] It is easy to understand that the input of LSTM is usually time series data. Therefore, to match the input format of the model. In this embodiment, it is necessary to perform data preprocessing on the real-time traffic data, convert it into a time series data set, and then input the time series data set into the long short-term memory network for network state analysis to obtain the time series data set.

[0062] In a feasible implementation manner, in this embodiment, step A1 may include steps A11 to A12:

[0063] Step A11: Normalize the real-time traffic data to obtain standard traffic data;

[0064] Step A12: Divide the standard traffic data according to a preset window length to generate a time series data set.

[0065] It can be understood that in this embodiment, the original real-time traffic data collected can be first normalized to obtain standard traffic data that eliminates the dimension difference. Then, in this embodiment, the set data input length of LSTM, that is, the above-mentioned preset window length, can be used as a time window to divide the standard traffic data to obtain a time series data set.

[0066] It should be noted that, in order to improve the accuracy of network state analysis of the LSTM network, in this embodiment, a self-attention mechanism can be added between LSTM layers to strengthen the feature extraction of key time windows (i.e., time periods with large traffic fluctuation amplitudes); meanwhile, the LSTM network in this embodiment can include a dual-branch structure, which are respectively used for short-term fluctuation or long-term trend analysis, and feature concatenation is performed based on multi-scale network features to obtain richer traffic features.

[0067] In addition, in order to further improve the adaptability of the subsequent generated dynamic congestion threshold, this embodiment can also perform lightweight improvement on the LSTM network. Specifically, this embodiment can use TCN (Temporal Convolutional Network) to replace some LSTM layers to reduce the computational complexity, and thus reduce the analysis time-consuming of LSTM.

[0068] It is easy to understand that after obtaining the dynamic congestion threshold in this embodiment, it can be sent to the switching chip through the API for real-time network state control, and the optimized explicit congestion notification marking is executed.

[0069] In this embodiment, aiming at the resource waste (such as buffer vacancy) or congestion deterioration (such as frequent packet loss due to too low threshold) caused by the existing ECN static threshold, this embodiment can deeply mine the temporal features through a long short-term memory network, and can combine the attention mechanism to strengthen the key features, dynamically generate the most suitable congestion threshold capture in the real-time state of the network, and perform optimized explicit congestion notification marking based on the dynamic congestion threshold, so as to significantly improve the adaptability of the optimized explicit congestion notification marking in the dynamic network environment.

[0070] In a feasible implementation manner, referring to Figure 2 , Figure 2 which is the second process schematic diagram of the first embodiment of the explicit congestion notification dynamic optimization method based on the long short-term memory network of this application, in this embodiment, after step S30, it further includes:

[0071] Step S40, obtaining the marked network performance data;

[0072] Step S50, iteratively optimizing the long short-term memory network based on the network performance data and a preset loss function.

[0073] It is easy to understand that the prior art can only passively wait for the network management and control of the ECN static threshold, and the network traffic pattern changes with time, user behavior, and application type. In order to improve the adaptability of the ECN marking, this embodiment can further adaptively modify the long short-term memory network according to the continuous evolution of the adaptive network topology and traffic pattern.

[0074] Therefore, in this embodiment, the network performance data corresponding to network control can be monitored in real time through the display congestion notification mark optimized by the dynamic congestion threshold, and the long short-term memory network can be iteratively optimized based on the feedback network performance data and the preset loss function to adapt to the network topology change.

[0075] In a feasible implementation manner, in this embodiment, step S50 includes steps S51 to S52:

[0076] Step S51, determining the current adjustment error based on the network performance data and the expected performance data;

[0077] Step S52, if it is detected that the duration of the deviation between the current adjustment error and the preset adjustment error exceeds the preset deviation duration, the long short-term memory network is iteratively optimized through the preset loss function.

[0078] It is easy to understand that the above network performance data in this embodiment may include the actual congestion situation, the change in packet loss rate, the real-time resource utilization rate, and the real-time end-to-end delay. In this embodiment, the current adjustment error can be determined based on the difference between it and the preset performance data.

[0079] If it is detected that the prediction error continues to be high, this embodiment can trigger the incremental training of the model. Specifically, it can be manifested as the duration when the current adjustment error is greater than the preset adjustment error or not within the error range corresponding to the preset adjustment error, that is, when the above deviation duration exceeds the preset deviation duration, this embodiment can iteratively optimize the long short-term memory network based on the preset loss function.

[0080] At this time, the formula of the above preset loss function can be expressed as follows:

[0081] L = min∑(ω1loss t + ω2R t + ω3L t ); (1)

[0082] In the formula, loss t is the current adjustment error, R t is the real-time resource utilization rate, L t is the real-time end-to-end delay, and ω1 to ω3 are the corresponding adjustment coefficients.

[0083] This embodiment proposes a method for dynamically optimizing explicit congestion notification parameters by combining long short-term memory network (LSTM) and explicit congestion notification (ECN). Compared with the traditional ECN marking method based on static trigger congestion marking, which cannot adapt to the dynamic changes of traffic, this embodiment can capture the long-term dependence relationship of time series data through the LSTM model, achieve high-precision prediction of network congestion status, effectively adapt to the complex and changeable network environment, and improve the dynamic adaptability;

[0084] In addition, this embodiment can also continuously optimize the long short-term memory network based on real-time network performance data to adapt to the changes of network topology and traffic patterns, and further improve the applicability of the method proposed in this embodiment.

[0085] This embodiment provides a method for dynamically optimizing explicit congestion notification based on long short-term memory network. The method includes: obtaining real-time network traffic data; normalizing the real-time traffic data to obtain standard traffic data; dividing the standard traffic data according to a preset window length to generate a time series data set; generating a dynamic congestion threshold through the long short-term memory network and the time series data set; performing explicit congestion notification marking based on the dynamic congestion threshold. Obtain the marked network performance data; determine the current adjustment error based on the network performance data and the expected performance data; if it is detected that the deviation duration between the current adjustment error and the preset adjustment error exceeds the preset deviation duration, the long short-term memory network is iteratively optimized through a preset loss function. This embodiment proposes a method for dynamically optimizing explicit congestion notification parameters by combining long short-term memory network (LSTM) and explicit congestion notification (ECN). Among them, this embodiment can capture the long-term dependence relationship of time series data through the LSTM model, achieve high-precision prediction of network congestion status, dynamically adjust the ECN threshold, effectively adapt to the complex and changeable network environment, and improve the dynamic adaptability. In addition, this embodiment can also continuously optimize the long short-term memory network based on real-time network performance data to adapt to the changes of network topology and traffic patterns, and further improve the applicability of the method proposed in this embodiment.

[0086] Based on the first embodiment of this application, in the second embodiment of this application, the same or similar content as the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated hereinafter.

[0087] On the basis of the first embodiment, please refer to Figure 3 , Figure 3 is a schematic flowchart of the second embodiment of the method for dynamically optimizing explicit congestion notification based on long short-term memory network in this application. In this embodiment, step A2 includes steps B1 to B2:

[0088] Step B1: Input the time series data set into a long short-term memory network for predicting the network congestion probability to obtain the dynamic congestion probability.

[0089] Step B2: Determine the dynamic congestion threshold according to the dynamic congestion probability.

[0090] It should be understood that traditional ECN has no prediction ability and can only respond to congestion passively, unable to prevent it in advance. Therefore, to further improve the adaptability in a dynamic network environment, this embodiment can analyze the temporal characteristics of network traffic data based on a long short-term memory network, thereby predicting the congestion probability of the current network, and then determining the real-time dynamic congestion threshold based on the predicted dynamic congestion probability.

[0091] At this time, the input of the long short-term memory network can be the historical traffic data of a preset window length (such as T seconds), and the output can be the probability of network congestion within the next N seconds, that is, the above-mentioned dynamic congestion probability. Subsequently, the congestion marking threshold (such as the queue length threshold) of ECN can be dynamically adjusted according to the LSTM prediction result. For example: when it is predicted that the real-time dynamic congestion probability exceeds 80%, the threshold is reduced to trigger the ECN marking in advance; when it is predicted that the real-time dynamic congestion probability is lower than 30%, the threshold is increased to delay the trigger of the ECN marking.

[0092] In a feasible implementation manner, in this embodiment, step B2 may include steps B21 to B22:

[0093] Step B21: Obtain the standard congestion threshold output by the long short-term memory network.

[0094] Step B22: Calculate the dynamic congestion threshold according to the standard congestion threshold and the dynamic congestion probability.

[0095] Threshold = Base_Threshold × (1 - α × P); (2)

[0096] In the formula, Threshold is the dynamic congestion threshold, Base_Threshold is the standard congestion threshold, P is the dynamic congestion probability, and α is the adjustment coefficient.

[0097] It is easy to understand that in this embodiment, the long short-term memory network can also output the recommended ECN marking threshold corresponding to the current network topology, that is, the above-mentioned standard congestion threshold. In the above formula for calculating the dynamic congestion probability, α is the adjustment coefficient, which can be used to control the threshold adjustment amplitude.

[0098] It can be understood that in the case of burst traffic, the threshold determined based on the predicted dynamic congestion probability in this embodiment can trigger the ECN marking 10 - 30 ms in advance, effectively avoiding queue overflow.

[0099] In this embodiment, to solve the problem that the existing ECN mechanism relies on a fixed threshold to trigger marking in the late stage of congestion, resulting in retransmission and delay accumulation, the dynamic congestion threshold can be determined by the network congestion probability predicted in real time by LSTM. Thus, while the threshold changes in real time with the network state, the ECN marking can be triggered further in advance, the network delay and packet loss can be reduced in a timely manner, and resource waste can be avoided.

[0100] This embodiment discloses inputting a time series data set into a long short-term memory network for predicting the network congestion probability to obtain the dynamic congestion probability; obtaining the standard congestion threshold output by the long short-term memory network; and calculating the dynamic congestion threshold according to the standard congestion threshold and the dynamic congestion probability. To solve the problem that the existing ECN mechanism relies on a fixed threshold to trigger marking in the late stage of congestion, resulting in retransmission and delay accumulation, the dynamic congestion threshold can be determined by the network congestion probability predicted in real time by LSTM. Thus, while the threshold changes in real time with the network state, the ECN marking can be triggered further in advance, the network delay and packet loss can be reduced in a timely manner, and resource waste can be avoided.

[0101] Exemplarily, to help understand the technical concept or technical principle of the explicit congestion notification dynamic optimization method based on the long short-term memory network after combining this embodiment with the above-mentioned Embodiment 1 and Embodiment 2, please refer to Figure 4 , Figure 4 which is a schematic flowchart of the brief process of the explicit congestion notification dynamic optimization method based on the long short-term memory network of this application, as follows:

[0102] Suppose the network of a certain data center is frequently congested due to traffic bursts and it is necessary to dynamically adjust the ECN (explicit congestion notification) threshold of the switch to avoid packet loss and delay. The explicit congestion notification dynamic optimization method based on the long short-term memory network proposed in this application aims to predict the congestion probability according to the real-time traffic and automatically optimize the ECN threshold. The operation process at this time can be as follows:

[0103] 1) Use sFlow to collect the real-time network traffic data (such as queue length, throughput, delay) of the switch port.

[0104] 2) Perform data preprocessing on the real-time network traffic data, normalize and slice the traffic data and then scale it to the interval [0, 1]. At this time, the sequence data can be divided according to a preset window length (such as 5 minutes) for LSTM training.

[0105] 3) Use the data set to perform offline training on the LSTM model. The LSTM prediction model can predict the queue length in the next 1 minute for congestion probability prediction.

[0106] 4) Deploy the trained offline model to the switch, dynamically adjust the ECN marking threshold according to the predicted congestion probability, and optimize the network condition according to the API and distribute it to the corresponding switching chip in the device.

[0107] 5) Monitor the network performance, and collect the feedback network metrics (packet loss rate, throughput) after adjustment in real time. If the prediction error continues to be high, trigger incremental training of the model through the feedback optimization module.

[0108] In summary, this application can significantly improve the accuracy of the ECN marking trigger timing by combining dynamic threshold adjustment, time series prediction enhancement, and closed-loop feedback optimization, and continuously optimize and update the LSTM model according to the evolution of the network topology and traffic pattern, thereby effectively improving the adaptability in the dynamic network environment.

[0109] It should be noted that the above examples are only used to understand this application, and do not constitute a limitation to the dynamic optimization method of explicit congestion notification based on the long short-term memory network of this application. Based on this technical concept, more forms of simple transformation are within the protection scope of this application.

[0110] This application also provides a dynamic optimization device for explicit congestion notification based on the long short-term memory network. Please refer to Figure 5 , Figure 5 which is the schematic diagram of the module structure of the dynamic optimization device for explicit congestion notification based on the long short-term memory network in the embodiment of this application. In this embodiment, the dynamic optimization device for explicit congestion notification based on the long short-term memory network includes:

[0111] A data acquisition module T1, which is used to acquire real-time network traffic data;

[0112] A data analysis module T2, which is used to generate a dynamic congestion threshold through the long short-term memory network and the real-time network traffic data;

[0113] An optimization module T3, which is used to perform explicit congestion notification marking based on the dynamic congestion threshold.

[0114] As a feasible implementation manner, in this embodiment, the data analysis module T2 is further used to perform data preprocessing on the real-time traffic data to generate a time series data set;

[0115] The data analysis module T2 is further used to generate a dynamic congestion threshold through the long short-term memory network and the time series data set.

[0116] As a feasible implementation manner, in this embodiment, the data analysis module T2 is further used to input the time series data set into the long short-term memory network to predict the network congestion probability and obtain the dynamic congestion probability;

[0117] The data analysis module T2 is also used to determine a dynamic congestion threshold according to the dynamic congestion probability.

[0118] As a feasible implementation manner, in this embodiment, the data analysis module T2 is also used to obtain the standard congestion threshold output by the long short-term memory network;

[0119] The data analysis module T2 is also used to calculate a dynamic congestion threshold according to the standard congestion threshold and the dynamic congestion probability;

[0120] Threshold = Base_Threshold × (1 - α × P);

[0121] In the formula, Threshold is the dynamic congestion threshold, Base_Threshold is the standard congestion threshold, P is the dynamic congestion probability, and α is an adjustment coefficient.

[0122] As a feasible implementation manner, in this embodiment, the data analysis module T2 is also used to normalize the real-time traffic data to obtain standard traffic data;

[0123] The data analysis module T2 is also used to divide the standard traffic data according to a preset window length to generate a time series data set.

[0124] As a feasible implementation manner, in this embodiment, the optimization module T3 is also used to obtain the marked network performance data;

[0125] The optimization module T3 is also used to iteratively optimize the long short-term memory network based on the network performance data and a preset loss function.

[0126] As a feasible implementation manner, in this embodiment, the optimization module T3 is also used to determine a current adjustment error based on the network performance data and the expected performance data;

[0127] The optimization module T3 is also used to, if it is detected that the deviation duration between the current adjustment error and a preset adjustment error exceeds a preset deviation duration, iteratively optimize the long short-term memory network through a preset loss function.

[0128] The explicit congestion notification dynamic optimization device based on a long short-term memory network provided by this application adopts the explicit congestion notification dynamic optimization method based on a long short-term memory network in the above embodiment, and can solve the technical problems of explicit congestion notification dynamic optimization based on a long short-term memory network. Compared with the prior art, the beneficial effects of the explicit congestion notification dynamic optimization device based on a long short-term memory network provided by this application are the same as those of the explicit congestion notification dynamic optimization method based on a long short-term memory network provided by the above embodiment, and other technical features in the explicit congestion notification dynamic optimization device based on a long short-term memory network are the same as the features disclosed in the method of the above embodiment, which will not be elaborated here.

[0129] This application provides an explicit congestion notification dynamic optimization device based on a long short-term memory network. The explicit congestion notification dynamic optimization device based on a long short-term memory network includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the explicit congestion notification dynamic optimization method in Embodiment 1 above.

[0130] Reference is made below Figure 6 , which shows a schematic structural diagram of an explicit congestion notification dynamic optimization device suitable for implementing the embodiments of this application. The explicit congestion notification dynamic optimization device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The explicit congestion notification dynamic optimization device shown is merely an example and should not impose any limitation on the functions and usage scope of the embodiments of this application.

[0131] As Figure 6As shown, the explicit congestion notification dynamic optimization device based on long short-term memory network may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM: Read Only Memory) 1002 or the program loaded from the storage device 1003 into the random access memory (RAM: Random Access Memory) 1004. In the random access memory 1004, various programs and data required for the operation of the explicit congestion notification dynamic optimization device based on long short-term memory network are also stored. The processing device 1001, the read-only memory 1002, and the random access memory 1004 are connected to each other through a bus 1005. The input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems may be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the explicit congestion notification dynamic optimization device based on long short-term memory network to communicate with other devices wirelessly or wiredly to exchange data. Although the figure shows an explicit congestion notification dynamic optimization device with various systems based on long short-term memory network, it should be understood that it is not required to implement or have all the shown systems. More or fewer systems may be implemented or had alternatively.

[0132] Specifically, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include an explicit congestion notification dynamic optimization program product based on long short-term memory network, which includes an explicit congestion notification dynamic optimization program carried on a computer-readable medium, and the explicit congestion notification dynamic optimization program contains program codes for performing the methods shown in the flowcharts. In such an embodiment, the explicit congestion notification dynamic optimization program can be downloaded and installed from the network through the communication device, or installed from the storage device 1003, or installed from the read-only memory 1002. When the explicit congestion notification dynamic optimization program is executed by the processing device 1001, the above functions defined in the methods of the embodiments disclosed in the present application are executed.

[0133] The explicit congestion notification dynamic optimization device based on long short-term memory network provided by this application adopts the explicit congestion notification dynamic optimization method based on long short-term memory network in the above-mentioned embodiment, and can solve the technical problems of explicit congestion notification dynamic optimization based on long short-term memory network. Compared with the prior art, the beneficial effects of the explicit congestion notification dynamic optimization device based on long short-term memory network provided by this application are the same as those of the explicit congestion notification dynamic optimization method based on long short-term memory network provided by the above-mentioned embodiment, and other technical features in the explicit congestion notification dynamic optimization device based on long short-term memory network are the same as the features disclosed in the method of the previous embodiment, and will not be elaborated here.

[0134] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0135] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0136] This application provides a storage medium with computer-readable program instructions stored thereon (i.e., the explicit congestion notification dynamic optimization program based on long short-term memory network), and the computer-readable program instructions are used to execute the explicit congestion notification dynamic optimization method in the above-mentioned embodiment.

[0137] The storage medium provided by this application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination of the above. More specific examples of the storage medium may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory, optical fibers, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this embodiment, the storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, device, or component. The program code contained on the storage medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0138] The above storage medium can be included in the explicit congestion notification dynamic optimization device based on long short-term memory networks; it can also exist separately without being assembled into the explicit congestion notification dynamic optimization device based on long short-term memory networks.

[0139] The above storage medium carries one or more programs. When the above one or more programs are executed by the explicit congestion notification dynamic optimization device based on long short-term memory networks, the explicit congestion notification dynamic optimization device based on long short-term memory networks is enabled to perform explicit congestion notification dynamic optimization based on long short-term memory networks.

[0140] The long short-term memory network-based explicit congestion notification dynamic optimization program code for performing the operations of this application can be written in one or more programming languages or combinations thereof. The above-mentioned programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0141] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of the systems, methods, and long short-term memory network-based explicit congestion notification dynamic optimization program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and this module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0142] The modules involved in the embodiments of this application can be implemented in software or in hardware. Among them, the name of the module does not constitute a limitation to the unit itself in some cases.

[0143] The readable storage medium provided by this application is a storage medium that stores computer-readable program instructions for executing the above-mentioned explicit congestion notification dynamic optimization method based on long short-term memory networks (i.e., the explicit congestion notification dynamic optimization program based on long short-term memory networks), and can solve the technical problems of explicit congestion notification dynamic optimization based on long short-term memory networks. Compared with the prior art, the beneficial effects of the storage medium provided by this application are the same as those of the explicit congestion notification dynamic optimization method based on long short-term memory networks provided in the above embodiments, and will not be elaborated here.

[0144] The above are only partial embodiments of this application, and do not limit the patent scope of this application. Any equivalent structural transformation made under the technical concept of this application by using the content of the specification and drawings of this application, or any direct / indirect application in other related technical fields, is included in the patent protection scope of this application.

Claims

1. An explicit congestion notification dynamic optimization method based on a long short-term memory network, characterized in that The method includes: Obtaining real-time network traffic data; Generating a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data; Performing a display congestion notification mark based on the dynamic congestion threshold.

2. The explicit congestion notification dynamic optimization method based on a long short-term memory network according to claim 1, wherein The step of generating a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data includes: Performing data preprocessing on the real-time traffic data to generate a time series data set; Generating a dynamic congestion threshold through a long short-term memory network and the time series data set.

3. The explicit congestion notification dynamic optimization method based on a long short-term memory network according to claim 2, wherein The step of generating a dynamic congestion threshold through a long short-term memory network and the time series data set includes: Inputting the time series data set into a long short-term memory network to predict the network congestion probability and obtaining a dynamic congestion probability; Determining a dynamic congestion threshold according to the dynamic congestion probability.

4. The explicit congestion notification dynamic optimization method based on long short-term memory network according to claim 3, wherein The step of determining a dynamic congestion threshold according to the dynamic congestion probability includes: Obtaining a standard congestion threshold output by the long short-term memory network; Calculating a dynamic congestion threshold according to the standard congestion threshold and the dynamic congestion probability; Threshold = Base_Threshold × (1 - α × P); In the formula, Threshold is the dynamic congestion threshold, Base_Threshold is the standard congestion threshold, P is the dynamic congestion probability, and α is an adjustment coefficient.

5. The explicit congestion notification dynamic optimization method based on a long short-term memory network according to claim 2, wherein The step of performing data preprocessing on the real-time traffic data to generate a time series data set includes: Normalizing the real-time traffic data to obtain standard traffic data; Dividing the standard traffic data according to a preset window length to generate a time series data set.

6. The explicit congestion notification dynamic optimization method based on long short-term memory network according to claim 1, characterized in that After performing a display congestion notification mark based on the dynamic congestion threshold, it further includes: Obtaining the marked network performance data; Iteratively optimizing the long short-term memory network based on the network performance data and a preset loss function.

7. The explicit congestion notification dynamic optimization method based on long short-term memory network according to claim 6, characterized in that, The step of iteratively optimizing the long short-term memory network based on the network performance data and a preset loss function includes: Determining a current adjustment error based on the network performance data and expected performance data; If it is detected that the deviation duration between the current adjustment error and a preset adjustment error exceeds a preset deviation duration, then iteratively optimizing the long short-term memory network through a preset loss function.

8. An explicit congestion notification dynamic optimization device based on a long short-term memory network, characterized in that, The explicit congestion notification dynamic optimization device based on a long short-term memory network includes: A data acquisition module for obtaining real-time network traffic data; A data analysis module for generating a dynamic congestion threshold through a long short-term memory network and the real-time network traffic data; An optimization module for performing a display congestion notification mark based on the dynamic congestion threshold.

9. An explicit congestion notification dynamic optimization device based on a long short-term memory network, characterized in that The explicit congestion notification dynamic optimization device based on a long short-term memory network includes: a memory, a processor, and an explicit congestion notification dynamic optimization program based on a long short-term memory network stored on the memory and executable on the processor, and the explicit congestion notification dynamic optimization program based on a long short-term memory network is configured to implement the steps of the explicit congestion notification dynamic optimization method according to any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium is a storage medium, and an explicit congestion notification dynamic optimization program based on a long short-term memory network is stored on the storage medium. When the explicit congestion notification dynamic optimization program based on the long short-term memory network is executed by a processor, the steps of the explicit congestion notification dynamic optimization method based on the long short-term memory network according to any one of claims 1 to 7 are implemented.

Images