A computer network management method

Abstract

The present invention relates to the technical field of computer network management, and in particular to a computer network management method, comprising the following steps: S1: real-time collection of network topology change data; S2: identification of potential bottleneck links and high-load areas; S3: generation of a traffic balancing strategy based on the traffic analysis results in S2; S4: dynamic adjustment of the transmission path of network data packets; S5: monitoring nodes monitor the implementation effect in real time to adjust the traffic balancing strategy in real time to cope with changes in network topology; S6: learning historical traffic data and feedback data in S5, and gradually adjusting and improving the traffic balancing strategy. The present invention, by implementing a dynamic traffic balancing strategy and machine learning technology, effectively improves the network's ability to adapt to changes, reduces congestion, and optimizes performance, thereby significantly improving the stability and operational efficiency of the network system.

Description

Technical Field

[0001] The present invention relates to the technical field of computer network management, and in particular to a computer network management method. Background Art

[0002] With the rapid development of information technology, network systems have become increasingly complex and large, especially in data centers, cloud computing platforms and large enterprise networks. The efficiency and accuracy of network traffic management are crucial to the stability and performance of the entire network. Traditional network traffic management methods rely on static routing configuration and manual intervention. These methods are often unable to cope with the growing amount of data and rapidly changing network conditions. In addition, traditional methods have obvious limitations in dealing with changes in network topology, link fault recovery and scheduling during peak traffic periods. The existence of these problems makes the network system prone to a series of problems such as high latency, uneven bandwidth utilization and reduced service quality.

[0003] Faced with the above problems, existing technologies have failed to provide an effective solution that can dynamically and intelligently adjust traffic distribution to adapt to changes in network status. Existing traffic management systems usually lack flexibility and cannot reflect the latest status of the network in real time, resulting in the inability to effectively prevent and handle network congestion in actual operations. Especially in large-scale networks, this situation will seriously affect the network's operational efficiency and user experience.

[0004] Therefore, there is an urgent need to develop a new technical approach to support more efficient and automated network traffic management, especially in environments with dense traffic and frequent changes in network topology. Summary of the invention

[0005] Based on the above purpose, the present invention provides a computer network management method.

[0006] A computer network management method comprises the following steps:

[0007] S1: Through the deployed distributed monitoring nodes, the network topology change data, including node connection status, link bandwidth and delay information, is collected in real time, and the data is transmitted to the central control module for processing;

[0008] S2: The central control module evaluates the traffic load of each link based on the network topology data collected in S1 and uses a preset multi-dimensional traffic analysis algorithm to identify potential bottleneck links and high-load areas;

[0009] S3: Generates a traffic balancing strategy based on the traffic analysis results in S2. The strategy covers processing methods for different traffic scenarios, including load redistribution, path switching, and traffic rate limiting.

[0010] S4: According to the traffic balancing strategy generated in S3, the transmission path of the network data packet is dynamically adjusted to reconfigure the data traffic to a path with low latency and sufficient bandwidth;

[0011] S5: Apply the transmission path reconfigured in S4 to the actual network environment, monitor the implementation effect in real time through the monitoring node, and feed the monitoring data back to the central control module to adjust the traffic balancing strategy in real time to cope with changes in the network topology;

[0012] S6: The central control module uses a preset machine learning algorithm to learn historical traffic data and feedback data in S5, gradually adjusts and improves the traffic balancing strategy, and improves the accuracy and efficiency of future traffic management.

[0013] Optionally, the S1 specifically includes:

[0014] S11: Deploy monitoring devices with data collection functions at predetermined nodes of the network, each monitoring device is equipped with a high-precision clock synchronization module to ensure that data collection between different monitoring nodes has a consistent time reference;

[0015] S12: Using the built-in network protocol analysis module of the monitoring device, the connection status of the connected network node is detected in real time, including the online status of the node, the connection quality and the stability of the connection;

[0016] S13: The monitoring device continuously monitors the bandwidth utilization of the connected link through the embedded bandwidth monitoring module, including the current bandwidth, peak bandwidth and bandwidth change trend, and generates real-time bandwidth data;

[0017] S14: The monitoring device uses an integrated network delay measurement module and a timestamp technology based on clock synchronization to accurately measure the transmission delay of data packets on the link and obtain real-time data of the delay;

[0018] S15: The node connection status, link bandwidth and delay information collected in S12 to S14 are aggregated and encrypted through the built-in encryption transmission module, and then transmitted to the central control module in real time.

[0019] Optionally, the S2 specifically includes:

[0020] S21: After the central control module receives the network topology data from the distributed monitoring nodes, it first cleans and filters the data through the built-in data preprocessing module to remove redundant information and abnormal data, and then formats the data;

[0021] S22: using a preset multi-dimensional traffic analysis algorithm, based on the pre-processed network topology data, a traffic load model of each link is generated, including the bandwidth utilization, traffic change trend and delay of the current link, so as to provide a comprehensive evaluation of the load status of each link;

[0022] S23: Based on the generated traffic load model, calculate the load critical value of each link, and identify the bottleneck link that is close to or exceeds the critical value by comparing the actual load with the critical value, and mark it as a high-risk link;

[0023] S24: Through the spatial clustering function of the multi-dimensional traffic analysis algorithm, the links with high load are classified into the same area, the high-load areas in the network are identified, and the traffic propagation trend of the area is predicted based on the relationship between the links;

[0024] S25: Summarize the bottleneck links and high-load area information identified in S23 and S24, generate a traffic load evaluation report, and transmit the evaluation results to the traffic balancing strategy generation module.

[0025] Optionally, the S22 specifically includes:

[0026] S221: Calculate the bandwidth utilization of each link using the following formula: Among them, U represents the bandwidth utilization of the link, B u Indicates the current bandwidth used by the link, B t Indicates the total bandwidth of the link;

[0027] S222: Calculate the traffic change rate of the link. The formula is: Among them, ΔF represents the traffic change rate of the link, f c Indicates the traffic of the link in the current time period, F p represents the traffic of the link in the previous time period, and T represents the time difference between the two time periods;

[0028] S223: Based on the delay information, calculate the average delay of each link: Where D is the average delay of the link, d j represents the delay value of the link at the jth measurement, and n represents the total number of measurements;

[0029] S224: The bandwidth utilization, traffic change rate and average delay data calculated in S221 to S223 are integrated to generate a comprehensive traffic load index of the link using a weighted model, and the formula is: L = w U ×U+w F ×ΔF+w D ×D, where L represents the comprehensive traffic load index of the link, and w U 、w F 、wD They represent the weight coefficients of bandwidth utilization, traffic change rate and delay respectively. The comprehensive traffic load index of each link is calculated by the above formula to form a complete traffic load model.

[0030] Optionally, the S23 specifically includes:

[0031] S231: Based on the traffic load model, calculate the load critical value for each link, the formula is: L c =α×L max , where L c represents the load critical value of the link, α is the critical coefficient preset by the system, and L max is the maximum carrying capacity of the link;

[0032] S232: Calculate the current load of each link to obtain the actual load value, the formula is: L a =w U ×U+w F ×ΔF+w D ×D, where L a represents the actual load value of the link, U, ΔF, and D represent bandwidth utilization, traffic change rate, and delay respectively, and w U ,w F 、w D is the corresponding weight coefficient;

[0033] S233: The actual load value L of the link a The corresponding load critical value L c For comparison, the formula is: ΔL=L a -L c , where ΔL represents the difference between the actual load value and the critical value. If ΔL ≥ 0, it means that the link has approached or exceeded its load critical value, and the link is identified as a bottleneck link;

[0034] S234: For the identified bottleneck link, mark it as a high-risk link.

[0035] Optionally, the S24 specifically includes:

[0036] S241: Based on the geographical location, connection relationship and traffic load value of the link, a spatial clustering algorithm is used to group links with similar geographical locations and similar load characteristics into a clustering unit. The objective function of the spatial clustering algorithm is: Among them, K is the number of clusters, C k is the kth clustering unit, d ik represents the geographical distance between the link and the cluster center k, is the actual load value of the link, is the average load value of cluster center k, β is the load influence factor. By minimizing the above objective function, high-load and geographically close links are clustered together to form a high-load area;

[0037] S242: Analyze the clustering results. If the average load value of the clustering unit is Exceeding the preset high load threshold L h , then the cluster unit is identified as a high-load area;

[0038] S243: For the identified high-load area, the connection relationship between links and historical traffic data are used to predict the future traffic propagation trend using a time series analysis algorithm. The prediction formula is: Among them, P t+1 is the expected flow rate in the high-load area at time t+1, P t is the current traffic value, γ is the traffic propagation influencing factor, and They are the actual load values ​​for the current and previous time periods respectively.

[0039] Optionally, the S3 specifically includes:

[0040] S31: Generation of load redistribution strategy. Based on the bottleneck links and high-load areas identified in S2, the load distribution adjustment amount of each link is calculated through the load redistribution algorithm. The formula is: Where, ΔL r Represents the load adjustment of each link, L c is the load critical value of the link, L a is the current actual load value, N is the number of links that can bear the adjusted load;

[0041] S32: Generation of path switching strategy. For the bottleneck link identified in S2, the central control module generates a path switching strategy and reallocates the transmission path of the data packet through the path optimization algorithm. Specifically, the shortest path priority algorithm is used. The formula is: Among them, P new is the optimized transmission path, d i is the distance of the first link on the path, D i is the link delay, L a is the actual load value of the link, w D and w L are the weight coefficients of delay and load respectively;

[0042] S33: Generation of traffic speed limit strategy. For high-load links that cannot be effectively alleviated by load redistribution or path switching, the central control module generates a traffic speed limit strategy and calculates the speed limit value of each link. The formula is: R limit =Lc -L a , where R limit is the speed limit of the link, L c is the load critical value of the link, L a is the current actual load value.

[0043] Optionally, the S4 specifically includes:

[0044] S41: According to the comprehensive traffic balancing strategy generated in S3, the priority of path selection is set to give priority to paths with low latency and sufficient bandwidth. The path priority calculation formula is:

[0045] Among them, P priority represents the priority of the path, D represents the delay of the path, and B a represents the available bandwidth of the path, w L and w B are the weight coefficients for delay and bandwidth respectively;

[0046] S42: selecting an optimal path for the network data packet to be transmitted according to the path priority in S41;

[0047] S43: During the data packet transmission process, the central control module monitors the load of the selected path in real time;

[0048] S44: If it is detected during the transmission process that the load of the path exceeds a critical value, the central control module will immediately re-evaluate the priorities of other alternative paths and dynamically adjust the transmission path of the data packet to the new priority path.

[0049] Optionally, the S5 specifically includes:

[0050] S51: receiving the path load and performance data fed back from the monitoring node, parsing and archiving the monitoring data, and generating a real-time monitoring report of the current network status, including the current load value, delay, bandwidth usage of each link, and whether there is a bottleneck link or a high-load area;

[0051] S52: according to the real-time monitoring report generated in S51, a trigger condition is set, and when the load of a certain link exceeds a preset critical value, or when a significant change in the network topology is detected, a dynamic adjustment process of the traffic balancing strategy is automatically triggered;

[0052] S53: Regeneration of the traffic balancing strategy. Based on the new monitoring data and trigger conditions, the traffic balancing strategy generation process in S3 is re-executed to calculate new load redistribution, path switching and traffic speed limit strategies.

[0053] Optionally, the S6 specifically includes:

[0054] S61: preprocessing the historical traffic data and the feedback data obtained in S5, including data cleaning, normalization and feature extraction. The extracted features include bandwidth utilization, delay, traffic change rate of the link, and indicators of network performance changes before and after each policy adjustment;

[0055] S62: using a preset machine learning algorithm, including a random forest, a support vector machine, or a neural network, to train the historical traffic data using the extracted features as input to generate a preliminary prediction model;

[0056] S63: After the initial model training is completed, the model is applied to a portion of historical data for verification and compared with actual feedback data. By analyzing the error between the prediction and the actual results, the model parameters are adjusted and the algorithm weights are corrected;

[0057] S64: After the model training and verification are completed, the traffic balancing strategy is regenerated based on the optimized machine learning model.

[0058] Beneficial effects of the present invention:

[0059] The present invention adopts a dynamic traffic balancing strategy, combined with real-time monitoring and data analysis. This method can adjust the traffic path according to the real-time status of the network, effectively disperse the pressure of high-load links, and reduce the delay caused by network congestion, thereby improving the overall performance and stability of the network. In addition, the introduced machine learning algorithm enables the system to learn from historical data and predict possible traffic patterns and network changes in the future, make adjustments in advance, and avoid potential network problems.

[0060] The present invention can significantly reduce network operating costs and improve network service quality by optimizing traffic distribution and enhancing the network's self-regulation capabilities. The promotion and use of this technology will bring significant economic benefits and social value to cloud computing platforms, data centers and other network-intensive industries, especially showing great potential in improving the efficiency of network resource utilization and ensuring network service reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0062] Figure 1 A schematic diagram of a computer network management method according to an embodiment of the present invention;

[0063] Figure 2The figure is a schematic diagram of the flow of collecting network topology structure change data according to an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0065] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0066] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0067] like Figure 1-Figure 2 As shown, a computer network management method comprises the following steps:

[0068] S1: Through the deployed distributed monitoring nodes, the network topology change data, including node connection status, link bandwidth and delay information, is collected in real time, and the data is transmitted to the central control module for processing;

[0069] S2: The central control module evaluates the traffic load of each link based on the network topology data collected in S1 and uses a preset multi-dimensional traffic analysis algorithm to identify potential bottleneck links and high-load areas;

[0070] S3: Generates traffic balancing strategies based on the traffic analysis results in S2. The strategies cover processing methods for different traffic scenarios, including load redistribution, path switching, and traffic speed limiting, to achieve global traffic balancing.

[0071] S4: According to the traffic balancing strategy generated in S3, the transmission path of network data packets is dynamically adjusted to reconfigure data traffic to a path with low latency and sufficient bandwidth to ensure load balancing on each link;

[0072] S5: Apply the transmission path reconfigured in S4 to the actual network environment, monitor the implementation effect in real time through the monitoring node, and feed the monitoring data back to the central control module to adjust the traffic balancing strategy in real time to cope with changes in the network topology;

[0073] S6: The central control module uses a preset machine learning algorithm to learn historical traffic data and feedback data in S5, gradually adjusts and improves the traffic balancing strategy, and improves the accuracy and efficiency of future traffic management.

[0074] S1 specifically includes:

[0075] S11: Deploy monitoring devices with data collection functions at predetermined nodes of the network. Each monitoring device is equipped with a high-precision clock synchronization module to ensure that data collection between different monitoring nodes has a consistent time reference to ensure data synchronization;

[0076] S12: Detect the connection status of the connected network nodes in real time through the built-in network protocol analysis module of the monitoring device, including the online status, connection quality and connection stability of the nodes, and transmit the information to the central control module;

[0077] S13: The monitoring device continuously monitors the bandwidth utilization of the connected link, including the current bandwidth, peak bandwidth and bandwidth change trend, through the embedded bandwidth monitoring module, generates real-time bandwidth data, and transmits it to the central control module;

[0078] S14: The monitoring device uses an integrated network delay measurement module and a timestamp technology based on clock synchronization to accurately measure the transmission delay of the data packet on the link, obtain the real-time data of the delay, and synchronize the delay data to the central control module;

[0079] S15: The node connection status, link bandwidth and delay information collected in S12 to S14 are aggregated and encrypted through the built-in encryption transmission module, and then transmitted to the central control module in real time to ensure the integrity and security of the data; through the specific deployment of the above-mentioned distributed monitoring nodes and data collection steps, accurate monitoring of changes in the network topology structure can be achieved, ensuring that the collected data is highly accurate and real-time, providing a solid data foundation for subsequent traffic analysis and strategy generation, and effectively improving the overall efficiency and reliability of network management.

[0080] S2 specifically includes:

[0081] S21: After the central control module receives the network topology data from the distributed monitoring nodes, it first cleans and filters the data through the built-in data preprocessing module to remove redundant information and abnormal data, and performs data formatting to ensure the accuracy of subsequent analysis;

[0082] S22: using a preset multi-dimensional traffic analysis algorithm, based on the pre-processed network topology data, a traffic load model of each link is generated, including the bandwidth utilization, traffic change trend and delay of the current link, so as to provide a comprehensive evaluation of the load status of each link;

[0083] S23: Based on the generated traffic load model, calculate the load critical value of each link, and identify the bottleneck link that is close to or exceeds the critical value by comparing the actual load with the critical value, and mark it as a high-risk link for subsequent optimization processing;

[0084] S24: Through the spatial clustering function of the multi-dimensional traffic analysis algorithm, the links with high load are classified into the same area, the high-load areas in the network are identified, and the traffic propagation trend of the area is predicted based on the relationship between the links, so as to formulate the corresponding load balancing strategy;

[0085] S25: Summarize the bottleneck links and high-load areas identified in S23 and S24, generate a traffic load assessment report, and transmit the assessment results to the traffic balancing strategy generation module as a basis for subsequent strategy generation; through the specific implementation method in the above step S2, the bottleneck links and high-load areas in the network can be accurately identified, which provides a scientific basis for the formulation of traffic balancing strategies, significantly improves the efficiency and reliability of network traffic management, and effectively avoids network congestion and performance degradation.

[0086] The traffic load model of each link generated in S22 specifically includes:

[0087] S221: Calculate the bandwidth utilization of each link using the following formula: Among them, U represents the bandwidth utilization of the link, B u Indicates the current bandwidth used by the link, B t Represents the total bandwidth of the link. The real-time bandwidth utilization of each link is obtained through the above calculation and used as the basic data of the traffic load model;

[0088] S222: Calculate the traffic change rate of the link. The formula is: Among them, ΔF represents the traffic change rate of the link, F c Indicates the traffic of the link in the current time period, F prepresents the traffic of the link in the previous time period, and T represents the time difference between the two time periods. The change trend of the link traffic is calculated through this formula to provide dynamic analysis data for the traffic load model;

[0089] S223: Based on the delay information, calculate the average delay of each link: Where D is the average delay of the link, d j represents the delay value of the link at the jth measurement, n represents the total number of measurements, and the transmission delay level of each link is evaluated by calculating the average delay and integrated into the traffic load model;

[0090] S224: The bandwidth utilization, traffic change rate and average delay data calculated in S221 to S223 are integrated to generate a comprehensive traffic load index of the link using a weighted model, and the formula is: L = w U ×U+w F ×ΔF+w D ×D, where L represents the comprehensive traffic load index of the link, and w U 、w F 、w D The weight coefficients respectively represent the bandwidth utilization, traffic change rate and delay. The comprehensive traffic load index of each link is calculated by the above formula to form a complete traffic load model. Through the specific algorithm and calculation formula in the above step S22, the traffic load model of each link can be accurately generated, which comprehensively considers multi-dimensional data such as bandwidth utilization, traffic change rate and delay, provides detailed and accurate basic information for traffic management, and can more effectively perform network optimization and load balancing.

[0091] S23 specifically includes:

[0092] S231: Based on the traffic load model, calculate the load critical value for each link, the formula is: L c =α×L max , where L c represents the load critical value of the link, α is the critical coefficient preset by the system, usually set according to the tolerance of the network, L max is the maximum carrying capacity of the link (i.e., the highest load value that the link can withstand under ideal conditions);

[0093] S232: Calculate the current load of each link to obtain the actual load value, the formula is: L a =w U ×U+w F ×ΔF+w D ×D, where L a represents the actual load value of the link, U, ΔF, and D represent bandwidth utilization, traffic change rate, and delay respectively, and w U ,wF 、w D is the corresponding weight coefficient, and this formula is used to calculate the comprehensive actual load of the link to ensure consistency with the traffic load model;

[0094] S233: The actual load value L of the link a The corresponding load critical value L c For comparison, the formula is: ΔL=L a -L c , where ΔL represents the difference between the actual load value and the critical value. If ΔL ≥ 0, it means that the link has approached or exceeded its load critical value, and the link is identified as a bottleneck link;

[0095] S234: For the identified bottleneck link, mark it as a high-risk link, and record its position in the network topology and its related load information so that it can be given priority in subsequent policy adjustments; through the above S23 step, the load critical value of each link can be accurately calculated, and by comparing the actual load with the critical value in real time, the bottleneck link that may be congested can be timely identified and marked as high-risk. This process provides reliable basic data for subsequent traffic balancing strategies, significantly improving the operating efficiency and stability of the network.

[0096] S24 specifically includes:

[0097] S241: Based on the geographical location, connection relationship and traffic load value of the link, a spatial clustering algorithm is used to group links with similar geographical locations and similar load characteristics into a clustering unit. The objective function of the spatial clustering algorithm is: Among them, K is the number of clusters, C k is the kth clustering unit, d ik represents the geographical distance between the link and the cluster center k, is the actual load value of the link, is the average load value of cluster center k, β is the load influence factor. By minimizing the above objective function, high-load and geographically close links are clustered together to form a high-load area;

[0098] S242: Analyze the clustering results. If the average load value of the clustering unit is Exceeding the preset high load threshold L h , the cluster unit is identified as a high-load area; the formula is: is the high load area, where L h A high-load area identification threshold preset for the system, and areas exceeding this threshold will be marked as high-load areas;

[0099] S243: For the identified high-load area, the connection relationship between links and historical traffic data are used to predict the future traffic propagation trend using a time series analysis algorithm. The prediction formula is: Among them, P t+1 is the expected flow rate in the high-load area at time t+1, P t is the current traffic value, γ is the traffic propagation influencing factor, and They are the actual load values ​​for the current and previous time periods, respectively. Through this prediction formula, the system can determine the future flow change trend in high-load areas in advance and provide a basis for flow control.

[0100] S3 specifically includes:

[0101] S31: Generation of load redistribution strategy. Based on the bottleneck links and high-load areas identified in S2, the load distribution adjustment amount of each link is calculated through the load redistribution algorithm. The formula is: Among them, ΔL r Represents the load adjustment of each link, L c is the load critical value of the link, L a is the current actual load value, and N is the number of links that can bear the load adjustment. Through this algorithm, the system can calculate the load that needs to be redistributed and evenly distribute it to other links to ensure the overall load balance of the network.

[0102] S32: Generation of path switching strategy. For the bottleneck link identified in S2, the central control module generates a path switching strategy and reallocates the transmission path of the data packet through the path optimization algorithm. Specifically, the shortest path first (SPF) algorithm is used. The formula is: Among them, P new is the optimized transmission path, d i is the distance of the first link on the path, D i is the link delay, L a is the actual load value of the link, w D and w L are the weight coefficients of delay and load respectively. Through the path switching strategy, data traffic is directed to a better path to reduce the load pressure of the bottleneck link;

[0103] S33: Generation of traffic speed limit strategy. For high-load links that cannot be effectively alleviated by load redistribution or path switching, the central control module generates a traffic speed limit strategy and calculates the speed limit value of each link. The formula is: R limit =L c -L a , where R limit is the speed limit of the link, L cis the load critical value of the link, L a The current actual load value is used. Through the speed limit strategy, the system limits the data flow entering the high-load link to avoid link overload. Through the above S3 step, it can generate an accurate traffic balancing strategy based on the traffic analysis results, including load redistribution, path switching, traffic speed limit and other processing methods to ensure the stable operation of the network. This strategy effectively reduces the load of the network bottleneck link and improves the overall network performance and response speed.

[0104] S4 specifically includes:

[0105] S41: According to the comprehensive traffic balancing strategy generated in S3, the priority of path selection is set to give priority to paths with low latency and sufficient bandwidth. The path priority calculation formula is:

[0106] Among them, P priority represents the priority of the path, D represents the delay of the path, and B a represents the available bandwidth of the path, w L and w B are the weight coefficients of delay and bandwidth respectively. The higher the path priority, the more likely it is to be selected for transmitting data packets;

[0107] S42: selecting an optimal path for the network data packet to be transmitted according to the path priority in S41;

[0108] S43: During the data packet transmission process, the central control module monitors the load of the selected path in real time to ensure that the path load does not exceed a preset load critical value;

[0109] S44: If the load of the path is detected to exceed the critical value during the transmission process, the central control module will immediately re-evaluate the priority of other alternative paths and dynamically adjust the transmission path of the data packet to the new priority path to avoid the data packet being stranded or delayed on the high-load path; through the specific technical solution in the above S4 step, the transmission path of the network data packet can be dynamically adjusted according to the real-time traffic balancing strategy to ensure that the data traffic is always transmitted through the path with the lowest latency and sufficient bandwidth, optimize network performance and reduce transmission delay. This dynamic adjustment mechanism improves the network's responsiveness and adaptability, and helps maintain the efficient operation of the network.

[0110] S5 specifically includes:

[0111] S51: receiving the path load and performance data fed back from the monitoring node, parsing and archiving the monitoring data, and generating a real-time monitoring report of the current network status, including the current load value, delay, bandwidth usage of each link, and whether there is a bottleneck link or a high-load area;

[0112] S52: According to the real-time monitoring report generated in S51, a trigger condition is set. When the load of a link exceeds a preset critical value, or when a significant change in the network topology is detected, the dynamic adjustment process of the traffic balancing strategy is automatically triggered. The trigger condition formula is:

[0113]

[0114] Among them, T adjust is the trigger signal for strategy adjustment, L a is the current link load value, L c is the critical load value of the link, when T adjust =1, it means that strategy adjustment is needed;

[0115] S53: Regeneration of traffic balancing strategy. Based on the new monitoring data and trigger conditions, the traffic balancing strategy generation process in S3 is re-executed to calculate new load redistribution, path switching and traffic speed limit strategies to ensure that the network can still maintain efficient operation under the new topology and load conditions.

[0116] S6 specifically includes:

[0117] S61: preprocess the historical traffic data and the feedback data obtained in S5, including data cleaning, normalization and feature extraction. The extracted features include link bandwidth utilization, latency, traffic change rate and indicators of network performance changes before and after each policy adjustment. By extracting these features, the system can capture the core factors that affect the effect of the traffic balancing strategy.

[0118] S62: Using preset machine learning algorithms, including random forests, support vector machines, or neural networks, the extracted features are used as input to train historical traffic data and generate a preliminary prediction model. By continuously optimizing model parameters and selecting the optimal algorithm configuration, the system can improve the accuracy of predicting future network traffic trends and policy effects;

[0119] S63: After the initial model training is completed, the model is applied to a portion of historical data for verification and compared with actual feedback data. By analyzing the error between the prediction and the actual results, the model parameters are adjusted and the algorithm weights are corrected to reduce the error and improve the prediction accuracy of the model;

[0120] S64: After model training and verification are completed, the traffic balancing strategy is regenerated based on the optimized machine learning model. At this time, the strategy will take into account more accurate future traffic prediction results. The system will make adjustments to potential network bottlenecks in advance based on the output results of the model, thereby improving the efficiency of traffic management.

[0121] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0122] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A computer network management method, It is characterized in that The following steps are involved: S1: Through the deployed distributed monitoring nodes, the network topology change data, including node connection status, link bandwidth and delay information, is collected in real time, and the data is transmitted to the central control module for processing; S2: The central control module evaluates the traffic load of each link based on the network topology data collected in S1 and uses a preset multi-dimensional traffic analysis algorithm to identify potential bottleneck links and high-load areas; S3: Generates a traffic balancing strategy based on the traffic analysis results in S2. The strategy covers processing methods for different traffic scenarios, including load redistribution, path switching, and traffic rate limiting. S4: According to the traffic balancing strategy generated in S3, the transmission path of the network data packet is dynamically adjusted to reconfigure the data traffic to a path with low latency and sufficient bandwidth; S5: Apply the transmission path reconfigured in S4 to the actual network environment, monitor the implementation effect in real time through the monitoring node, and feed the monitoring data back to the central control module to adjust the traffic balancing strategy in real time to cope with changes in the network topology; S6: The central control module uses a preset machine learning algorithm to learn historical traffic data and feedback data in S5, gradually adjusts and improves the traffic balancing strategy, and improves the accuracy and efficiency of future traffic management.

2. A computer network management method according to claim 1, It is characterized in that The S1 specifically includes: S11: Deploy monitoring devices with data collection functions at predetermined nodes of the network, each monitoring device is equipped with a high-precision clock synchronization module to ensure that data collection between different monitoring nodes has a consistent time reference; S12: Using the built-in network protocol analysis module of the monitoring device, the connection status of the connected network node is detected in real time, including the online status of the node, the connection quality and the stability of the connection; S13: The monitoring device continuously monitors the bandwidth utilization of the connected link through the embedded bandwidth monitoring module, including the current bandwidth, peak bandwidth and bandwidth change trend, and generates real-time bandwidth data; S14: The monitoring device uses an integrated network delay measurement module and a timestamp technology based on clock synchronization to accurately measure the transmission delay of data packets on the link and obtain real-time data of the delay; S15: The node connection status, link bandwidth and delay information collected in S12 to S14 are aggregated and encrypted through the built-in encryption transmission module, and then transmitted to the central control module in real time.

3. A computer network management method according to claim 1, It is characterized in that The S2 specifically includes: S21: After the central control module receives the network topology data from the distributed monitoring nodes, it first cleans and filters the data through the built-in data preprocessing module to remove redundant information and abnormal data, and then formats the data; S22: using a preset multi-dimensional traffic analysis algorithm, based on the pre-processed network topology data, a traffic load model of each link is generated, including the bandwidth utilization, traffic change trend and delay of the current link, so as to provide a comprehensive evaluation of the load status of each link; S23: Based on the generated traffic load model, calculate the load critical value of each link, and identify the bottleneck link that is close to or exceeds the critical value by comparing the actual load with the critical value, and mark it as a high-risk link; S24: Through the spatial clustering function of the multi-dimensional traffic analysis algorithm, the links with high load are classified into the same area, the high-load areas in the network are identified, and the traffic propagation trend of the area is predicted based on the relationship between the links; S25: Summarize the bottleneck links and high-load area information identified in S23 and S24, generate a traffic load evaluation report, and transmit the evaluation results to the traffic balancing strategy generation module.

4. A computer network management method according to claim 3, It is characterized in that The S22 specifically includes: S221: Calculate the bandwidth utilization of each link using the following formula: Among them, U represents the bandwidth utilization of the link, B u Indicates the current bandwidth used by the link, B t Indicates the total bandwidth of the link; S222: Calculate the traffic change rate of the link. The formula is: Among them, ΔF represents the traffic change rate of the link, F c Indicates the traffic of the link in the current time period, F p represents the traffic of the link in the previous time period, and T represents the time difference between the two time periods; S223: Based on the delay information, calculate the average delay of each link: Where D is the average delay of the link, d j represents the delay value of the link at the jth measurement, and n represents the total number of measurements; S224: The bandwidth utilization, traffic change rate and average delay data calculated in S221 to S223 are integrated to generate a comprehensive traffic load index of the link using a weighted model, and the formula is: L = w U ×U+w F ×ΔF+w D ×D, where L represents the comprehensive traffic load index of the link, and w U 、w F 、w D They represent the weight coefficients of bandwidth utilization, traffic change rate and delay respectively. The comprehensive traffic load index of each link is calculated by the above formula to form a complete traffic load model.

5. A computer network management method according to claim 4, It is characterized in that The S23 specifically includes: S231: Based on the traffic load model, calculate the load critical value for each link, the formula is: L c =α×L max , where L c represents the load critical value of the link, α is the critical coefficient preset by the system, and L max is the maximum carrying capacity of the link; S232: Calculate the current load of each link to obtain the actual load value, the formula is: L a =w U ×U+w F ×ΔF+w D ×D, where L a represents the actual load value of the link, U, ΔF, and D represent bandwidth utilization, traffic change rate, and delay respectively, and w U ,w F 、w D is the corresponding weight coefficient; S233: The actual load value L of the link a The corresponding load critical value L c For comparison, the formula is: ΔL=L a -L c , where ΔL represents the difference between the actual load value and the critical value. If ΔL ≥ 0, it means that the link has approached or exceeded its load critical value, and the link is identified as a bottleneck link; S234: For the identified bottleneck link, mark it as a high-risk link.

6. A computer network management method according to claim 5, It is characterized in that The S24 specifically includes: S241: Based on the geographical location, connection relationship and traffic load value of the link, a spatial clustering algorithm is used to group links with similar geographical locations and similar load characteristics into a clustering unit. The objective function of the spatial clustering algorithm is: Among them, K is the number of clusters, C k is the kth clustering unit, di k表 represents the geographical distance between the link and the cluster center k, is the actual load value of the link, is the average load value of cluster center k, β is the load influence factor. By minimizing the above objective function, high-load and geographically close links are clustered together to form a high-load area; S242: Analyze the clustering results. If the average load value of the clustering unit is Exceeding the preset high load threshold L h , then the cluster unit is identified as a high-load area; S243: For the identified high-load area, the connection relationship between links and historical traffic data are used to predict the future traffic propagation trend using a time series analysis algorithm. The prediction formula is: Among them, P t+1 is the expected flow rate in the high-load area at time t+1, P t is the current traffic value, γ is the traffic propagation influencing factor, and They are the actual load values ​​for the current and previous time periods respectively.

7. A computer network management method according to claim 1, It is characterized in that The S3 specifically includes: S31: Generation of load redistribution strategy. Based on the bottleneck links and high-load areas identified in S2, the load distribution adjustment amount of each link is calculated through the load redistribution algorithm. The formula is: Where, ΔL r Represents the load adjustment of each link, L c is the load critical value of the link, L a is the current actual load value, N is the number of links that can bear the adjusted load; S32: Generation of path switching strategy. For the bottleneck link identified in S2, the central control module generates a path switching strategy and reallocates the transmission path of the data packet through the path optimization algorithm. Specifically, the shortest path priority algorithm is used. The formula is: Among them, P new is the optimized transmission path, d i is the distance of the first link on the path, D i is the link delay, L a is the actual load value of the link, w D and w L are the weight coefficients of delay and load respectively; S33: Generation of traffic speed limit strategy. For high-load links that cannot be effectively alleviated by load redistribution or path switching, the central control module generates a traffic speed limit strategy and calculates the speed limit value of each link. The formula is: R limit =L c -L a , where R limit is the speed limit of the link, L c is the load critical value of the link, L a is the current actual load value.

8. A computer network management method according to claim 1, It is characterized in that The S4 specifically includes: S41: According to the comprehensive traffic balancing strategy generated in S3, the priority of path selection is set to give priority to paths with low latency and sufficient bandwidth. The path priority calculation formula is: Among them, P priority represents the priority of the path, D represents the delay of the path, and B a represents the available bandwidth of the path, w L and w B are the weight coefficients for delay and bandwidth respectively; S42: selecting an optimal path for the network data packet to be transmitted according to the path priority in S41; S43: During the data packet transmission process, the central control module monitors the load of the selected path in real time; S44: If it is detected during the transmission process that the load of the path exceeds a critical value, the central control module will immediately re-evaluate the priorities of other alternative paths and dynamically adjust the transmission path of the data packet to the new priority path.

9. A computer network management method according to claim 1, It is characterized in that The S5 specifically includes: S51: receiving the path load and performance data fed back from the monitoring node, parsing and archiving the monitoring data, and generating a real-time monitoring report of the current network status, including the current load value, delay, bandwidth usage of each link, and whether there is a bottleneck link or a high-load area; S52: according to the real-time monitoring report generated in S51, a trigger condition is set, and when the load of a certain link exceeds a preset critical value, or when a significant change in the network topology is detected, a dynamic adjustment process of the traffic balancing strategy is automatically triggered; S53: Regeneration of the traffic balancing strategy. Based on the new monitoring data and trigger conditions, the traffic balancing strategy generation process in S3 is re-executed to calculate new load redistribution, path switching and traffic speed limit strategies.

10. A computer network management method according to claim 1, It is characterized in that The S6 specifically includes: S61: preprocessing the historical traffic data and the feedback data obtained in S5, including data cleaning, normalization and feature extraction. The extracted features include bandwidth utilization, delay, traffic change rate of the link, and indicators of network performance changes before and after each policy adjustment. S62: using a preset machine learning algorithm, including a random forest, a support vector machine, or a neural network, to train the historical traffic data using the extracted features as input to generate a preliminary prediction model; S63: After the initial model training is completed, the model is applied to a portion of historical data for verification and compared with actual feedback data. By analyzing the error between the prediction and the actual results, the model parameters are adjusted and the algorithm weights are corrected; S64: After the model training and verification are completed, the traffic balancing strategy is regenerated based on the optimized machine learning model.

Images