Internet of Things intelligent monitoring method and system

Abstract

The invention provides an intelligent monitoring method and system for the Internet of Things. The intelligent monitoring method for the Internet of Things comprises the following steps: monitoring data processing load parameters of network node clusters corresponding to servers included in the Internet of Things in real time; wherein each server corresponds to one network node cluster; when the data processing load parameter of the network node cluster exceeds a preset load parameter threshold value, obtaining a correlation coefficient between network nodes contained in the network node cluster and other network node clusters; determining a target network node by using the correlation coefficient between the network node and other network node clusters; and transferring the target network node to a server to which the network node cluster corresponding to the maximum correlation coefficient of the target network node belongs. The system comprises modules corresponding to the steps of the method.

Description

Technical Field

[0001] The present invention provides an Internet of Things intelligent monitoring method and system, belonging to the technical field of Internet of Things monitoring. Background Art

[0002] In today's digital age, the Internet of Things (IoT) has been widely adopted, encompassing numerous fields such as intelligent transportation, smart homes, and industrial monitoring. IoT systems connect and collaborate with numerous sensors, devices, and servers to collect, transmit, and process data. With the continuous expansion of IoT and the increasing complexity of its application scenarios, the amount of data generated within the system has exploded. Each server needs to process massive amounts of data, and its corresponding network node cluster bears the heavy burden of data processing. However, existing technologies for balancing server loads have many shortcomings.

[0003] On the one hand, many IoT systems lack effective real-time monitoring mechanisms, making it difficult to accurately and timely obtain the data processing load parameters of network node clusters. This makes it impossible to predict server load in advance. When the server load is too high, data processing delays, data loss, and even server crashes may occur, seriously affecting the stability and reliability of the system.

[0004] On the other hand, even when detecting excessive server load, existing load balancing strategies are often not intelligent or flexible enough. Traditional methods may simply allocate tasks without fully considering the relationships between network nodes and other node clusters. This makes it impossible to achieve optimal resource allocation during load adjustment, effectively alleviate server pressure, and fully utilize the resources of the entire IoT system.

[0005] Therefore, there is an urgent need for an IoT intelligent monitoring method that can monitor server load in real time and accurately, and perform intelligent load balancing based on the correlation between network nodes, so as to improve the performance, stability and resource utilization efficiency of the IoT system. Summary of the Invention

[0006] The present invention provides an Internet of Things intelligent monitoring method and system to solve the technical problems existing in the above-mentioned prior art. The technical solutions adopted are as follows:

[0007] An Internet of Things intelligent monitoring method, the Internet of Things intelligent monitoring method comprising:

[0008] Real-time monitoring of the data processing load parameters of the network node cluster corresponding to each server included in the Internet of Things;

[0009] When the data processing load parameter of the network node cluster exceeds a preset load parameter threshold, obtaining the correlation coefficient between the network node included in the network node cluster and other network node clusters;

[0010] Determining a target network node using a correlation coefficient between the network node and other network node clusters;

[0011] The target network node is transferred to a server to which a network node cluster corresponding to the maximum value of the target network node association coefficient belongs.

[0012] Furthermore, the data processing load parameters of the network node cluster corresponding to each server included in the Internet of Things are monitored in real time, including:

[0013] Real-time monitoring of the operating parameters of each network node included in the network node cluster corresponding to each server included in the Internet of Things; wherein the operating parameters of the network node include CPU utilization, memory utilization and cache hit rate;

[0014] Real-time monitoring of the percentage change in CPU utilization and memory utilization corresponding to each percentage unit increase in cache hit rate; wherein the value range of the percentage unit corresponding to the cache hit rate is 0.1%-0.3%;

[0015] Obtaining the resource fluctuation load coefficient corresponding to each network node using the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase of the cache hit rate;

[0016] The data processing load parameter of the network node cluster is obtained by utilizing the resource fluctuation load coefficient corresponding to each network node included in the network node cluster.

[0017] Furthermore, the data processing load parameter of the network node cluster is obtained by utilizing the resource fluctuation load coefficient corresponding to each network node included in the network node cluster, including:

[0018] Retrieving the resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0019] Obtaining a standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster according to the resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0020] The data processing load parameter of the network node cluster is obtained by utilizing the resource fluctuation load coefficient corresponding to each network node included in the network node cluster and the standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster.

[0021] Furthermore, when the data processing load parameter of the network node cluster exceeds a preset load parameter threshold, obtaining the correlation coefficient between the network node included in the network node cluster and other network node clusters includes:

[0022] When the data processing load parameter of the network node cluster exceeds the preset load parameter threshold, the amount of data exchanged between the network nodes contained in the network node cluster and other network node clusters per unit time is retrieved.

[0023] The correlation coefficient of each network node with other network node clusters is obtained by utilizing the data volume of data exchanged between the network nodes included in the network node cluster and other network node clusters per unit time.

[0024] Furthermore, determining the target network node by using the correlation coefficient between the network node and other network node clusters includes:

[0025] The contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which the network node is located is obtained by using the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located;

[0026] Normalizing the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located so that the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is projected to the interval (0, 1);

[0027] Performing ratio processing on the data processing load parameter of the network node cluster and a preset load parameter threshold value to obtain a ratio between the preset load parameter threshold value and the data processing load parameter of the network node cluster as a ratio parameter;

[0028] Obtaining a load disturbance coefficient of the data processing load of the network node cluster to which each network node belongs by using the contribution intensity coefficient and the ratio parameter of each network node to the data processing load parameter of the network node cluster to which it belongs;

[0029] The network node corresponding to the maximum value of the load disturbance coefficient is used as the target network node.

[0030] An Internet of Things intelligent monitoring system, the Internet of Things intelligent monitoring system comprising:

[0031] A monitoring module for real-time monitoring of the data processing load parameters of the network node clusters corresponding to each server included in the Internet of Things; a correlation coefficient acquisition module for acquiring the correlation coefficients of the network nodes included in the network node cluster with other network node clusters when the data processing load parameter of the network node cluster exceeds a preset load parameter threshold;

[0032] a target network node determination module, configured to determine a target network node by using a correlation coefficient between the network node and other network node clusters;

[0033] The node switching module is used to switch the target network node to the server to which the network node cluster corresponding to the maximum value of the correlation coefficient belongs.

[0034] Furthermore, the monitoring module includes:

[0035] A first parameter monitoring module is used to monitor in real time the operating parameters of each network node included in the network node cluster corresponding to each server included in the Internet of Things; wherein the operating parameters of the network node include CPU utilization, memory utilization and cache hit rate;

[0036] The second parameter monitoring module is used to monitor in real time the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase in cache hit rate; wherein the value range of the percentage unit corresponding to the cache hit rate is 0.1%-0.3%;

[0037] A resource fluctuation load coefficient acquisition module is used to obtain the resource fluctuation load coefficient corresponding to each network node by using the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase of the cache hit rate;

[0038] The data processing load parameter acquisition module is used to acquire the data processing load parameter of the network node cluster by using the resource fluctuation load coefficient corresponding to each network node included in the network node cluster.

[0039] Furthermore, the data processing load parameter acquisition module includes:

[0040] A resource fluctuation load coefficient retrieving module is used to retrieve the resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0041] A standard deviation acquisition module, configured to acquire a standard deviation of a resource fluctuation load coefficient corresponding to a network node cluster according to a resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0042] The load parameter acquisition execution module is used to obtain the data processing load parameter of the network node cluster by using the resource fluctuation load coefficient corresponding to each network node included in the network node cluster and the resource fluctuation load coefficient standard deviation corresponding to the network node cluster.

[0043] Furthermore, the correlation coefficient acquisition module includes:

[0044] An interactive data volume retrieving module, configured to retrieve the data volume of data interaction between the network nodes included in the network node cluster and other network node clusters per unit time when the data processing load parameter of the network node cluster exceeds a preset load parameter threshold;

[0045] The correlation coefficient acquisition execution module is used to obtain the correlation coefficient of each network node with other network node clusters by using the data exchange data volume between the network nodes included in the network node cluster and other network node clusters per unit time.

[0046] Furthermore, the target network node determination module includes:

[0047] A contribution intensity coefficient acquisition module is used to obtain the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which the network node is located by using the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located;

[0048] a normalization processing module, configured to perform normalization processing on a contribution intensity coefficient of each network node to a data processing load parameter of the network node cluster in which it is located, so that the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is projected to an interval of (0, 1);

[0049] a ratio parameter acquisition module, configured to perform ratio processing on the data processing load parameter of the network node cluster and a preset load parameter threshold value, and obtain a ratio between the preset load parameter threshold value and the data processing load parameter of the network node cluster as a ratio parameter;

[0050] a load disturbance coefficient acquisition module, configured to acquire a load disturbance coefficient of each network node to the data processing load of the network node cluster in which it is located by using a contribution intensity coefficient and a ratio parameter of each network node to the data processing load parameter of the network node cluster in which it is located;

[0051] The target network node screening module is used to select the network node corresponding to the maximum value of the load disturbance coefficient as the target network node.

[0052] Beneficial effects of the present invention:

[0053] The present invention proposes an intelligent monitoring method and system for the Internet of Things to achieve load balancing between servers and network node clusters in the Internet of Things. By switching out nodes in overloaded node clusters, the workload of different servers and node clusters is made more even, avoiding performance degradation or even crash of some servers due to excessive load, and improving the stability and reliability of the entire Internet of Things system. The ownership of network nodes is reasonably allocated, the processing power of each server is fully utilized, and idle resources or excessively concentrated use of resources are avoided, thereby improving the utilization efficiency of system resources and reducing operating costs. It effectively alleviates the data processing pressure of network node clusters, reduces data processing delays and waiting time, speeds up data processing, improves the overall data processing performance and response speed of the Internet of Things system, and provides users with a smoother service experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A flow chart of the method of the present invention;

[0055] Figure 2 This is a system block diagram of the system of the present invention. DETAILED DESCRIPTION

[0056] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0057] The embodiment of the present invention proposes an Internet of Things intelligent monitoring method, such as Figure 1 As shown, the IoT intelligent monitoring method includes:

[0058] Real-time monitoring of the data processing load parameters of the network node clusters corresponding to each server included in the Internet of Things; each server corresponds to a network node cluster;

[0059] When the data processing load parameter of the network node cluster exceeds a preset load parameter threshold, obtaining the correlation coefficient between the network node included in the network node cluster and other network node clusters;

[0060] Determining a target network node using a correlation coefficient between the network node and other network node clusters;

[0061] The target network node is transferred to a server to which a network node cluster corresponding to the maximum value of the target network node association coefficient belongs.

[0062] The working principle of the above technical solution is as follows: the method first monitors the data processing load parameters of the network node clusters corresponding to each server in the Internet of Things in real time. This is the starting point of the entire process. Through continuous monitoring, the current data processing workload of each network node cluster is obtained. The monitored data processing load parameters are compared with a preset load parameter threshold. When the data processing load parameter of a network node cluster exceeds the threshold, it means that the current processing pressure of this node cluster is too high and needs to be adjusted. When the network node cluster is determined to be overloaded, the correlation coefficient between the network nodes contained in this network node cluster and other network node clusters is obtained. The correlation coefficient measures the closeness of the connection between the node and other node clusters. Based on the obtained correlation coefficient, the target network node is determined from the overloaded network node cluster. The determination method may be based on a certain rule, such as selecting nodes with specific correlation coefficient characteristics (such as value size, distribution, etc.) as target nodes. The determined target network node is transferred to the server belonging to the network node cluster with the highest correlation coefficient. Through this transfer, some processing tasks in the originally overloaded node cluster are transferred to other relatively idle servers, achieving load balancing.

[0063] The effect of the above technical solution is to achieve load balancing between servers and network node clusters in the Internet of Things. By transferring nodes from overloaded node clusters, the workload of different servers and node clusters is more evenly distributed, preventing performance degradation or even crashes of some servers due to excessive load, and improving the stability and reliability of the entire Internet of Things system. It rationally allocates the ownership of network nodes, fully utilizes the processing power of each server, avoids idle resources or excessive concentration of resources, improves the utilization efficiency of system resources, and reduces operating costs. It effectively alleviates the data processing pressure of network node clusters, reduces data processing delays and waiting times, speeds up data processing, and improves the overall data processing performance and response speed of the Internet of Things system, providing users with a smoother service experience.

[0064] One embodiment of the present invention monitors data processing load parameters of network node clusters corresponding to various servers included in the Internet of Things in real time, including:

[0065] Real-time monitoring of the operating parameters of each network node included in the network node cluster corresponding to each server included in the Internet of Things; wherein the operating parameters of the network node include CPU utilization, memory utilization and cache hit rate;

[0066] Real-time monitoring of the percentage change in CPU utilization and memory utilization corresponding to each percentage unit increase in cache hit rate; wherein the value range of the percentage unit corresponding to the cache hit rate is 0.1%-0.3%;

[0067] Obtaining the resource fluctuation load coefficient corresponding to each network node using the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase of the cache hit rate;

[0068] The resource fluctuation load coefficient corresponding to each network node is obtained by the following formula:

[0069]

[0070] Where R represents the resource fluctuation load coefficient corresponding to each network node; α and β represent the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase in the cache hit rate of each network node; ΔH represents the unit change of cache hit rate; k represents the preset cache sensitivity coefficient, which ranges from 3 to 8; (1-e -k·ΔH ) preserves the amplification effect of cache changes while avoiding exponential explosion; the denominator retains ΔH to maintain the basic inverse proportional relationship, and the k value can dynamically adjust the sensitivity of different scenarios; the product of α and β reflects the synergistic effect of CPU and memory, which conforms to the complex characteristics of real loads.

[0071] The data processing load parameter of the network node cluster is obtained by utilizing the resource fluctuation load coefficient corresponding to each network node included in the network node cluster.

[0072] The working principle of the above technical solution is to monitor the CPU utilization, memory utilization, and cache hit rate of each network node in the network node cluster corresponding to each server in the Internet of Things in real time. These parameters reflect the hardware resource usage of the network node during data processing and serve as the basis for subsequent analysis. Furthermore, the percentage change in CPU and memory utilization for each specific percentage increase in the cache hit rate (ranging from 0.1% to 0.3%) is monitored in real time. This is because changes in the cache hit rate affect CPU and memory usage, and monitoring this correlation provides a deeper understanding of the dynamic characteristics of network node resource usage. Based on the percentage change in CPU and memory utilization corresponding to the cache hit rate change, the resource fluctuation load coefficient corresponding to each network node is calculated. This coefficient comprehensively reflects the fluctuation in hardware resource usage when the cache hit rate changes, quantifying the degree of dynamic changes in network node resource usage. Using the resource fluctuation load coefficient of each network node in the network node cluster, a data processing load parameter for the entire network node cluster is calculated using a method (such as weighted average) to provide a comprehensive measure of the data processing load of the network node cluster.

[0073] The above technical solution achieves the following benefits: By comprehensively monitoring multiple operating parameters of network nodes and the dynamic correlations between these parameters, the data processing load of network nodes and node clusters can be more accurately assessed. Compared to monitoring a single operating parameter, this load assessment is more accurate and comprehensive, providing a reliable basis for subsequent decision-making. Real-time monitoring and calculation of resource fluctuation load factors can promptly detect abnormal fluctuations in network node resource usage, providing early warning of potential risks before network node clusters become overloaded. This allows for timely adjustment measures to prevent system performance degradation or even failure due to excessive load. Accurately understanding the data processing load parameters of network node clusters allows for more rational allocation of system resources, directing resources to less-loaded node clusters. This optimizes resource allocation across the entire IoT system, improves resource utilization efficiency, and ensures stable and efficient system operation. Furthermore, the resource fluctuation load factor calculation combines cache hit rate, CPU utilization, and memory utilization—interrelated parameters that reflect the operational status of network nodes from different perspectives. The cache hit rate reflects data read efficiency, while CPU and memory utilization reflect computing and storage resource usage. This comprehensive assessment of multiple parameters comprehensively captures the complex resource usage of network nodes during data processing. Compared to single-parameter evaluations, it more accurately reflects the actual performance and load of nodes, improving the accuracy of network node performance assessments. It also focuses on the dynamic impact of cache hit rate changes on CPU and memory utilization, capturing the percentage change. In actual operation, cache hit rate fluctuations trigger a chain reaction in CPU and memory resource usage. Capturing this dynamic correlation provides a deeper understanding of the underlying logic of resource usage, accurately grasping the resource fluctuation characteristics of network nodes under different operating conditions, and more accurately calculating the resource fluctuation load factor, providing a reliable basis for performance evaluation. By monitoring the changes in related parameters for every specific percentage increase in cache hit rate (0.1%-0.3%), it can keenly capture subtle fluctuations in network node resource usage. Even small changes in cache hit rate can be used to detect their potential impact on node performance through changes in related parameters, allowing for timely identification of node performance trends and early detection of performance issues, enabling early intervention and optimization. Using cache hit rate changes as triggers, the system updates CPU and memory utilization data in real time and calculates the resource fluctuation load factor. Once the operating status of a network node changes, such as fluctuations in cache hit rate due to changes in data access patterns, this information is quickly reflected in related parameters and the resource fluctuation load factor, enabling a rapid response to changes in network node performance and making the system more sensitive to performance fluctuations. Accurately calculating the resource fluctuation load factor allows for early identification of abnormal fluctuations in network node resource usage. When the resource fluctuation load factor shows an abnormality, it indicates potential risks to node performance, such as increased resource competition or processing capacity bottlenecks. This allows operations and maintenance personnel to intervene early and take measures such as adjusting cache strategies and optimizing resource allocation to prevent performance issues from worsening and ensure the stable operation of network nodes and the entire IoT system.Accurate resource fluctuation load coefficients provide strong support for load balancing across network node clusters. Based on this coefficient, the system determines node load and allocates tasks appropriately, shifting tasks from overloaded nodes to less loaded ones. This maintains a balanced load across nodes, preventing overloaded individual nodes from impacting overall system stability. This enhances the IoT system's ability to cope with load fluctuations and improves overall operational stability.

[0074] In one embodiment of the present invention, obtaining a data processing load parameter of a network node cluster by using a resource fluctuation load coefficient corresponding to each network node included in the network node cluster includes:

[0075] Retrieving the resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0076] Obtaining a standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster according to the resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0077] The data processing load parameter of the network node cluster is obtained by utilizing the resource fluctuation load coefficient corresponding to each network node included in the network node cluster and the standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster.

[0078] The data processing load parameter of the network node cluster is obtained by the following formula:

[0079]

[0080] Where L represents the data processing load parameter of the network node cluster; n represents the number of network nodes contained in the network node cluster; R i represents the resource fluctuation load coefficient corresponding to the i-th network node; R max Indicates the maximum value of resource fluctuation load coefficient corresponding to the network node; R p Represents the average value of resource fluctuation load coefficient corresponding to all network nodes in the network node cluster; σ R Indicates the standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster.

[0081] The working principle of the above technical solution is: first, the resource fluctuation load coefficient corresponding to each network node in the network node cluster is retrieved. These coefficients reflect the fluctuation of resource usage of each network node and are the basic data for subsequent calculations. According to the resource fluctuation load coefficient of each network node that has been retrieved, the standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster is calculated. The standard deviation is used to measure the degree of dispersion of the data. Here it can reflect the difference in the resource fluctuation load coefficients of each node in the network node cluster, that is, the degree of dispersion of the resource fluctuation of each node. Combined with the resource fluctuation load coefficient corresponding to each network node in the network node cluster and the calculated standard deviation, the data processing load parameter of the network node cluster is comprehensively obtained through a specific calculation method (such as weighted calculation, etc.). This parameter integrates the resource fluctuation of each node and its degree of dispersion, thereby comprehensively reflecting the data processing load status of the entire network node cluster.

[0082] The above technical solution not only considers the resource fluctuation load coefficient of each network node but also accounts for the dispersion of resource fluctuations between nodes through the standard deviation. Differences in resource fluctuations between nodes can affect the data processing capacity of the entire node cluster. Comprehensively considering these differences can more accurately reflect the actual data processing load of the network node cluster, avoiding the estimation bias caused by focusing solely on the coefficients of individual nodes. In actual operation, the loads of nodes within a network node cluster are not evenly distributed, and the standard deviation can reflect this imbalance. Incorporating this into the calculation makes the resulting data processing load parameter more accurate to the actual load of the network node cluster, providing a more reliable basis for system decision-making. The standard deviation of the resource fluctuation load coefficient can help identify nodes within the network node cluster experiencing abnormal fluctuations. A large standard deviation indicates significant differences in resource fluctuations between nodes, potentially indicating that some nodes may be overloaded or underloaded. The system can then make timely adjustments accordingly to prevent individual node anomalies from affecting the stability of the entire network node cluster and maintain stable system operation. Accurate network node cluster data processing load parameters facilitate more effective load balancing. Based on this parameter, the system can more accurately determine which node clusters are overloaded and which are relatively idle, thereby allocating tasks more scientifically, avoiding excessive load concentration in certain node clusters, and improving the stability of the entire IoT system. Standard deviation is sensitive to subtle changes in data and can capture small changes in resource fluctuations within network node clusters. Even if the resource fluctuation load coefficient of some nodes does not change much, if the degree of dispersion between nodes changes, the standard deviation will also change accordingly, which in turn affects the data processing load parameter, allowing the system to promptly perceive subtle changes in the performance of network node clusters. The data processing load parameter calculated based on standard deviation can quickly reflect dynamic changes in the performance of network node clusters. When the network environment or business needs change, the system can respond quickly based on this parameter and adjust the resource allocation strategy in a timely manner to ensure efficient system operation.

[0083] In one embodiment of the present invention, when the data processing load parameter of the network node cluster exceeds a preset load parameter threshold, obtaining the correlation coefficient between the network node included in the network node cluster and other network node clusters includes:

[0084] When the data processing load parameter of the network node cluster exceeds the preset load parameter threshold, the amount of data exchanged between the network nodes contained in the network node cluster and other network node clusters per unit time is retrieved.

[0085] The correlation coefficient of each network node with other network node clusters is obtained by utilizing the data volume of data exchanged between the network nodes included in the network node cluster and other network node clusters per unit time.

[0086] The correlation coefficient between each network node and other network node clusters is obtained by the following formula:

[0087]

[0088] Among them, G ij represents the correlation coefficient between the i-th network node and the j-th network node cluster; C ij represents the average amount of data interaction per unit time between the i-th network node and the j-th network node cluster; C zj Indicates the average amount of total data interaction per unit time of the jth network node cluster; C0 represents the dynamic baseline interaction volume (the median of the total system data interaction volume). The above formula measures the degree of interaction dependence between node i and node cluster j, through 1-e -Cij / C0 Achieve nonlinear saturation response to data volume.

[0089] The working principle of the above technical solution is as follows: First, the calculated data processing load parameter of a network node cluster is compared with a preset load parameter threshold. If the threshold is exceeded, it indicates that the network node cluster is facing significant data processing pressure, and further analysis of its relationship with other network node clusters is required to find possible solutions. Once the network node cluster is determined to be overloaded, the amount of data exchanged per unit time between each network node in the cluster and other network node clusters is retrieved. This step aims to collect data that reflects the closeness of connections between network nodes, as the amount of data exchange, to a certain extent, reflects the degree of association between different network node clusters. The acquired amount of data exchange per unit time is used to calculate the correlation coefficient between each network node and other network node clusters. Using a specific algorithm (such as normalization based on the amount of data exchange), the data exchange volume is converted into a numerical value that intuitively reflects the degree of association, namely the correlation coefficient. The higher the coefficient, the closer the corresponding network node is to the other network node clusters.

[0090] The above technical solution achieves the following: Using the volume of data exchanged per unit time between a network node and other network node clusters as the basis for calculating the correlation coefficient accurately reflects the actual degree of connection between different network nodes. Data exchange volume is a direct and key indicator of inter-node relationships. The correlation coefficient calculated based on this volume can more accurately assess the connections between network node clusters, providing accurate data support for subsequent decision-making. Different network nodes interact with other network node clusters in varying ways. By analyzing the volume of data exchanged per unit time, it is possible to effectively distinguish the varying degrees of connection between each network node and its cluster. This avoids simplistic generalizations and allows for a more nuanced understanding of the complex relationships within the network structure, helping to more effectively address the issue of overloaded network node clusters. In the IoT environment, data exchange between network nodes changes dynamically with evolving business needs and the network environment. Calculating the correlation coefficient based on the volume of data exchanged per unit time allows for timely adaptation to these changes. As the data traffic distribution within the network changes, the correlation coefficient adjusts accordingly, enabling the system to make informed decisions based on the latest network status, demonstrating strong flexibility and adaptability. Different IoT application scenarios have distinct data interaction patterns and characteristics. By focusing on the volume of data interaction per unit time, the correlation coefficient can be determined to adapt to diverse business scenarios. Whether for data-intensive applications or those with high real-time requirements, the correlation coefficient can be accurately calculated based on actual data interaction, providing strong support for network optimization in different business scenarios. By calculating the correlation coefficient, nodes with high correlation to currently overloaded network node clusters can be quickly located from among numerous network nodes. This helps the system quickly identify potential resource allocation targets, reducing the time and cost of finding solutions and improving the system's efficiency in addressing issues such as network congestion. Accurate correlation coefficients can help the system develop more optimized resource allocation strategies. By routing target network nodes to the server belonging to the corresponding network node cluster based on the maximum correlation coefficient, network resources can be more rationally allocated, avoiding blind resource adjustments. This improves resource utilization efficiency across the IoT system and ensures efficient operation.

[0091] In one embodiment of the present invention, determining a target network node using a correlation coefficient between the network node and other network node clusters includes:

[0092] The contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which the network node is located is obtained by using the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located;

[0093] The contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is obtained by the following formula:

[0094]

[0095] Where S represents the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located; R represents the resource fluctuation load coefficient corresponding to each network node; R p represents the average value of resource fluctuation load coefficients corresponding to all network nodes in the network node cluster; δ represents the preset smoothing factor, which is used to prevent numerical instability when the denominator approaches zero, and its value range is 0.1-0.3; L represents the data processing load parameter of the network node cluster; L y Indicates the preset load parameter threshold;

[0096] Normalizing the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located so that the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is projected to the interval (0, 1);

[0097] Performing ratio processing on the data processing load parameter of the network node cluster and a preset load parameter threshold value to obtain a ratio between the preset load parameter threshold value and the data processing load parameter of the network node cluster as a ratio parameter;

[0098] Obtaining a load disturbance coefficient of the data processing load of the network node cluster to which each network node belongs by using the contribution intensity coefficient and the ratio parameter of each network node to the data processing load parameter of the network node cluster to which it belongs;

[0099] The load disturbance coefficient of each network node to the data processing load of the network node cluster in which it is located is obtained by the following formula:

[0100]

[0101] Among them, P r represents the load disturbance coefficient of each network node to the data processing load of the network node cluster where it is located; S represents the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster where it is located; G max Indicates the maximum value of the correlation coefficient between a network node and other network node clusters; G avg Represents the average value of the correlation coefficient between the network node and other network node clusters; B LRepresents a ratio parameter. S (the contribution intensity coefficient of each network node to the data processing load parameter of its network node cluster) reflects the contribution of a single network node to the data processing load of its network node cluster. A larger contribution intensity coefficient indicates a greater impact on the cluster's load. It plays a more important role in the calculation of the load disturbance coefficient and is a fundamental parameter for measuring a node's contribution to the cluster's load. (the ratio of the maximum value to the average value of the correlation coefficient between a network node and other network node clusters) reflects the relative difference in the degree of correlation between a network node and other network node clusters. max With G avg The larger the ratio, the more significant the difference in the degree of association between the node and other node clusters, and the node may have special performance in data interaction. This ratio is combined with the contribution intensity coefficient to comprehensively evaluate the impact of the characteristics of the node in the network structure on the load. L (The ratio between the preset load parameter threshold and the data processing load parameter of the network node cluster) reflects the degree of deviation of the current network node cluster data processing load from the preset threshold. L It will amplify or reduce the effect of the combination of the contribution intensity coefficient and the correlation coefficient ratio. When the node cluster load far exceeds the threshold, B L The above formula combines these three factors. Based on the contribution intensity coefficient and the node network characteristics reflected by the correlation coefficient ratio, the load disturbance coefficient is adjusted based on the node cluster load deviation threshold. This ratio parameter is then used to determine the load disturbance coefficient, which comprehensively measures the impact of each network node on the data processing load of its cluster.

[0102] The network node corresponding to the maximum value of the load disturbance coefficient is used as the target network node.

[0103] The working principle of the above technical solution is as follows: First, the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located are comprehensively considered. A specific calculation method (such as weighted calculation) is used to derive the contribution intensity coefficient of each network node to the data processing load parameter of its network node cluster. This coefficient reflects the proportion or influence of each network node in the data processing load of the entire network node cluster. The calculated contribution intensity coefficient of each network node is normalized, and its numerical range is projected to the interval (0, 1). The purpose of normalization is to make the contribution intensity coefficients of different network nodes comparable and eliminate the influence caused by different original numerical ranges. The data processing load parameter of the network node cluster is compared with a preset load parameter threshold to obtain the ratio between the two, which is used as a ratio parameter. This ratio parameter can reflect the degree of deviation of the data processing load of the current network node cluster from the preset threshold. Combining the contribution intensity coefficient of each network node with the calculated ratio parameter, a specific algorithm (such as multiplication) is used to calculate the load disturbance coefficient of each network node to the data processing load of its network node cluster. The load perturbation coefficient comprehensively considers the node's own contribution and the load deviation of the entire node cluster. The load perturbation coefficients of all network nodes are compared, and the node with the largest load perturbation coefficient is identified as the target network node. Selecting the node with the largest load perturbation coefficient indicates that it has the greatest impact on the load of the network node cluster. Selecting it as the target node for subsequent processing (such as forwarding to a server belonging to another network node cluster) is expected to alleviate the current network node cluster's load pressure to the greatest extent.

[0104] The above technical solution achieves the following: By comprehensively considering multiple factors, including resource fluctuation load coefficients and network node cluster data processing load parameters, it accurately calculates each network node's contribution intensity coefficient and load disturbance coefficient to the load of its corresponding network node cluster. This allows the system to precisely locate the target network node with the greatest impact on the data processing load of the network node cluster, avoiding misselection and improving decision accuracy. The calculation of the load disturbance coefficient not only considers the node's own contribution but also incorporates the load deviation of the network node cluster, more accurately reflecting the actual impact of the network node on the load of the entire node cluster. This comprehensive approach ensures that the determined target network nodes better meet actual needs and provides a more reliable basis for subsequent load balancing and resource allocation. Selecting the network node with the highest load disturbance coefficient as the target network node can minimize the load pressure on the current network node cluster. Because this node has the greatest impact on the load, relocating it can quickly reduce the load of the network node cluster, improve system response speed, reduce data processing latency, and thus enhance the operational efficiency of the entire IoT system. Accurately determining the target network node allows for more targeted resource allocation. This avoids blind attempts to transfer different nodes, reduces ineffective operations during resource allocation, optimizes the resource allocation process, saves time and resources, and improves resource utilization efficiency. By rationally identifying the target network node and performing the transfer, the load distribution within the network node cluster is more evenly distributed, avoiding system instability caused by excessive load on individual nodes. This helps enhance the system's resistance to interference and improves its stability in the face of various network changes and load fluctuations. Effective load balancing and resource allocation can reduce the risk of overloading the network node cluster, reducing the possibility of system failures and service interruptions caused by overload, ensuring the continuous and stable operation of the IoT system and improving system reliability and availability.

[0105] On the other hand, the formula comprehensively considers the node's own contribution, its association characteristics with other node clusters, and the overall load deviation of the node cluster, comprehensively covering the key factors that affect the load disturbance of the network node to its node cluster. Compared with a single factor measurement, it can more accurately reflect the actual impact of the node on the load, provide a basis for accurately judging the role of the node on the network load, and improve the accuracy of load assessment. Ratio parameter B L The formula can be adaptively adjusted according to the different situations of network node cluster load relative to the threshold. Different load deviations from the threshold have different effects on the load disturbance coefficient calculation, making the calculation results more in line with the actual network operation status and further improving the accuracy. It can sensitively detect subtle differences in the degree of connectivity between a network node and other node clusters. Even if the inter-node connectivity changes slightly, this ratio can be detected and reflected in the load disturbance coefficient, making the system more sensitive to changes in node connectivity within the network structure. As the network operates, node contribution strength, connectivity, and cluster load dynamically change. The formula calculates the load disturbance coefficient based on these changes in real time, rapidly responding to changes in network status and promptly reflecting the dynamic changes in a node's impact on the load. Accurately calculating the load disturbance coefficient helps pinpoint nodes with the greatest impact on the load of network node clusters. The system can then rationally allocate resources, shift tasks from overloaded nodes, achieve load balancing, reduce system instability caused by uneven node load, and enhance system stability. By accurately assessing node load disturbances, the system can take proactive measures, such as adjusting task allocation and adding resources, before node or cluster loads reach dangerous levels, to prevent overload risks and ensure continued stable system operation.

[0106] An embodiment of the present invention proposes an Internet of Things intelligent monitoring system, such as Figure 2 As shown, the Internet of Things intelligent monitoring system includes:

[0107] A monitoring module is used to monitor in real time the data processing load parameters of the network node clusters corresponding to each server included in the Internet of Things; each server corresponds to a network node cluster;

[0108] a correlation coefficient acquisition module, configured to acquire correlation coefficients between the network nodes included in the network node cluster and other network node clusters when a data processing load parameter of the network node cluster exceeds a preset load parameter threshold;

[0109] a target network node determination module, configured to determine a target network node by using a correlation coefficient between the network node and other network node clusters;

[0110] The node switching module is used to switch the target network node to the server to which the network node cluster corresponding to the maximum value of the correlation coefficient belongs.

[0111] The working principle of the above technical solution is as follows: the method first monitors the data processing load parameters of the network node clusters corresponding to each server in the Internet of Things in real time. This is the starting point of the entire process. Through continuous monitoring, the current data processing workload of each network node cluster is obtained. The monitored data processing load parameters are compared with a preset load parameter threshold. When the data processing load parameter of a network node cluster exceeds the threshold, it means that the current processing pressure of this node cluster is too high and needs to be adjusted. When the network node cluster is determined to be overloaded, the correlation coefficient between the network nodes contained in this network node cluster and other network node clusters is obtained. The correlation coefficient measures the closeness of the connection between the node and other node clusters. Based on the obtained correlation coefficient, the target network node is determined from the overloaded network node cluster. The determination method may be based on a certain rule, such as selecting nodes with specific correlation coefficient characteristics (such as value size, distribution, etc.) as target nodes. The determined target network node is transferred to the server belonging to the network node cluster with the highest correlation coefficient. Through this transfer, some processing tasks in the originally overloaded node cluster are transferred to other relatively idle servers, achieving load balancing.

[0112] The effect of the above technical solution is to achieve load balancing between servers and network node clusters in the Internet of Things. By transferring nodes from overloaded node clusters, the workload of different servers and node clusters is more evenly distributed, preventing performance degradation or even crashes of some servers due to excessive load, and improving the stability and reliability of the entire Internet of Things system. It rationally allocates the ownership of network nodes, fully utilizes the processing power of each server, avoids idle resources or excessive concentration of resources, improves the utilization efficiency of system resources, and reduces operating costs. It effectively alleviates the data processing pressure of network node clusters, reduces data processing delays and waiting times, speeds up data processing, and improves the overall data processing performance and response speed of the Internet of Things system, providing users with a smoother service experience.

[0113] In one embodiment of the present invention, the monitoring module includes:

[0114] A first parameter monitoring module is used to monitor in real time the operating parameters of each network node included in the network node cluster corresponding to each server included in the Internet of Things; wherein the operating parameters of the network node include CPU utilization, memory utilization and cache hit rate;

[0115] The second parameter monitoring module is used to monitor in real time the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase in cache hit rate; wherein the value range of the percentage unit corresponding to the cache hit rate is 0.1%-0.3%;

[0116] A resource fluctuation load coefficient acquisition module is used to obtain the resource fluctuation load coefficient corresponding to each network node by using the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase of the cache hit rate;

[0117] The resource fluctuation load coefficient corresponding to each network node is obtained by the following formula:

[0118]

[0119] Where R represents the resource fluctuation load coefficient corresponding to each network node; α and β represent the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase in the cache hit rate of each network node; ΔH represents the unit change of cache hit rate; k represents the preset cache sensitivity coefficient, which ranges from 3 to 8; (1-e -k·ΔH ) preserves the amplification effect of cache changes while avoiding exponential explosion; the denominator retains ΔH to maintain the basic inverse proportional relationship, and the k value can dynamically adjust the sensitivity of different scenarios; the product of α and β reflects the synergistic effect of CPU and memory, which conforms to the complex characteristics of real loads.

[0120] The data processing load parameter acquisition module is used to acquire the data processing load parameter of the network node cluster by using the resource fluctuation load coefficient corresponding to each network node included in the network node cluster.

[0121] The working principle of the above technical solution is to monitor the CPU utilization, memory utilization, and cache hit rate of each network node in the network node cluster corresponding to each server in the Internet of Things in real time. These parameters reflect the hardware resource usage of the network node during data processing and serve as the basis for subsequent analysis. Furthermore, the percentage change in CPU and memory utilization for each specific percentage increase in the cache hit rate (ranging from 0.1% to 0.3%) is monitored in real time. This is because changes in the cache hit rate affect CPU and memory usage, and monitoring this correlation provides a deeper understanding of the dynamic characteristics of network node resource usage. Based on the percentage change in CPU and memory utilization corresponding to the cache hit rate change, the resource fluctuation load coefficient corresponding to each network node is calculated. This coefficient comprehensively reflects the fluctuation in hardware resource usage when the cache hit rate changes, quantifying the degree of dynamic changes in network node resource usage. Using the resource fluctuation load coefficient of each network node in the network node cluster, a data processing load parameter for the entire network node cluster is calculated using a method (such as weighted average) to provide a comprehensive measure of the data processing load of the network node cluster.

[0122] The above technical solution achieves the following benefits: By comprehensively monitoring multiple operating parameters of network nodes and the dynamic correlations between these parameters, the data processing load of network nodes and node clusters can be more accurately assessed. Compared to monitoring a single operating parameter, this load assessment is more accurate and comprehensive, providing a reliable basis for subsequent decision-making. Real-time monitoring and calculation of resource fluctuation load factors can promptly detect abnormal fluctuations in network node resource usage, providing early warning of potential risks before network node clusters become overloaded. This allows for timely adjustment measures to prevent system performance degradation or even failure due to excessive load. Accurately understanding the data processing load parameters of network node clusters allows for more rational allocation of system resources, directing resources to less-loaded node clusters. This optimizes resource allocation across the entire IoT system, improves resource utilization efficiency, and ensures stable and efficient system operation. Furthermore, the resource fluctuation load factor calculation combines cache hit rate, CPU utilization, and memory utilization—interrelated parameters that reflect the operational status of network nodes from different perspectives. The cache hit rate reflects data read efficiency, while CPU and memory utilization reflect computing and storage resource usage. This comprehensive assessment of multiple parameters comprehensively captures the complex resource usage of network nodes during data processing. Compared to single-parameter evaluations, it more accurately reflects the actual performance and load of nodes, improving the accuracy of network node performance assessments. It also focuses on the dynamic impact of cache hit rate changes on CPU and memory utilization, capturing the percentage change. In actual operation, cache hit rate fluctuations trigger a chain reaction in CPU and memory resource usage. Capturing this dynamic correlation provides a deeper understanding of the underlying logic of resource usage, accurately grasping the resource fluctuation characteristics of network nodes under different operating conditions, and more accurately calculating the resource fluctuation load factor, providing a reliable basis for performance evaluation. By monitoring the changes in related parameters for every specific percentage increase in cache hit rate (0.1%-0.3%), it can keenly capture subtle fluctuations in network node resource usage. Even small changes in cache hit rate can be used to detect their potential impact on node performance through changes in related parameters, allowing for timely identification of node performance trends and early detection of performance issues, enabling early intervention and optimization. Using cache hit rate changes as triggers, the system updates CPU and memory utilization data in real time and calculates the resource fluctuation load factor. Once the operating status of a network node changes, such as fluctuations in cache hit rate due to changes in data access patterns, this information is quickly reflected in related parameters and the resource fluctuation load factor, enabling a rapid response to changes in network node performance and making the system more sensitive to performance fluctuations. Accurately calculating the resource fluctuation load factor allows for early identification of abnormal fluctuations in network node resource usage. When the resource fluctuation load factor shows an abnormality, it indicates potential risks to node performance, such as increased resource competition or processing capacity bottlenecks. This allows operations and maintenance personnel to intervene early and take measures such as adjusting cache strategies and optimizing resource allocation to prevent performance issues from worsening and ensure the stable operation of network nodes and the entire IoT system.Accurate resource fluctuation load coefficients provide strong support for load balancing across network node clusters. Based on this coefficient, the system determines node load and allocates tasks appropriately, shifting tasks from overloaded nodes to less loaded ones. This maintains a balanced load across nodes, preventing overloaded individual nodes from impacting overall system stability. This enhances the IoT system's ability to cope with load fluctuations and improves overall operational stability.

[0123] In one embodiment of the present invention, the data processing load parameter acquisition module includes:

[0124] A resource fluctuation load coefficient retrieving module is used to retrieve the resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0125] A standard deviation acquisition module, configured to acquire a standard deviation of a resource fluctuation load coefficient corresponding to a network node cluster according to a resource fluctuation load coefficient corresponding to each network node included in the network node cluster;

[0126] The load parameter acquisition execution module is used to obtain the data processing load parameter of the network node cluster by using the resource fluctuation load coefficient corresponding to each network node included in the network node cluster and the resource fluctuation load coefficient standard deviation corresponding to the network node cluster.

[0127] The data processing load parameter of the network node cluster is obtained by the following formula:

[0128]

[0129] Where L represents the data processing load parameter of the network node cluster; n represents the number of network nodes contained in the network node cluster; R i represents the resource fluctuation load coefficient corresponding to the i-th network node; R max Indicates the maximum value of resource fluctuation load coefficient corresponding to the network node; R p Represents the average value of resource fluctuation load coefficient corresponding to all network nodes in the network node cluster; σ R Indicates the standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster.

[0130] The working principle of the above technical solution is: first, the resource fluctuation load coefficient corresponding to each network node in the network node cluster is retrieved. These coefficients reflect the fluctuation of resource usage of each network node and are the basic data for subsequent calculations. According to the resource fluctuation load coefficient of each network node that has been retrieved, the standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster is calculated. The standard deviation is used to measure the degree of dispersion of the data. Here it can reflect the difference in the resource fluctuation load coefficients of each node in the network node cluster, that is, the degree of dispersion of the resource fluctuation of each node. Combined with the resource fluctuation load coefficient corresponding to each network node in the network node cluster and the calculated standard deviation, the data processing load parameter of the network node cluster is comprehensively obtained through a specific calculation method (such as weighted calculation, etc.). This parameter integrates the resource fluctuation of each node and its degree of dispersion, thereby comprehensively reflecting the data processing load status of the entire network node cluster.

[0131] The above technical solution not only considers the resource fluctuation load coefficient of each network node but also accounts for the dispersion of resource fluctuations between nodes through the standard deviation. Differences in resource fluctuations between nodes can affect the data processing capacity of the entire node cluster. Comprehensively considering these differences can more accurately reflect the actual data processing load of the network node cluster, avoiding the estimation bias caused by focusing solely on the coefficients of individual nodes. In actual operation, the loads of nodes within a network node cluster are not evenly distributed, and the standard deviation can reflect this imbalance. Incorporating this into the calculation makes the resulting data processing load parameter more accurate to the actual load of the network node cluster, providing a more reliable basis for system decision-making. The standard deviation of the resource fluctuation load coefficient can help identify nodes within the network node cluster experiencing abnormal fluctuations. A large standard deviation indicates significant differences in resource fluctuations between nodes, potentially indicating that some nodes may be overloaded or underloaded. The system can then make timely adjustments accordingly to prevent individual node anomalies from affecting the stability of the entire network node cluster and maintain stable system operation. Accurate network node cluster data processing load parameters facilitate more effective load balancing. Based on this parameter, the system can more accurately determine which node clusters are overloaded and which are relatively idle, thereby allocating tasks more scientifically, avoiding excessive load concentration in certain node clusters, and improving the stability of the entire IoT system. Standard deviation is sensitive to subtle changes in data and can capture small changes in resource fluctuations within network node clusters. Even if the resource fluctuation load coefficient of some nodes does not change much, if the degree of dispersion between nodes changes, the standard deviation will also change accordingly, which in turn affects the data processing load parameter, allowing the system to promptly perceive subtle changes in the performance of network node clusters. The data processing load parameter calculated based on standard deviation can quickly reflect dynamic changes in the performance of network node clusters. When the network environment or business needs change, the system can respond quickly based on this parameter and adjust the resource allocation strategy in a timely manner to ensure efficient system operation.

[0132] In one embodiment of the present invention, the correlation coefficient acquisition module includes:

[0133] An interactive data volume retrieving module, configured to retrieve the data volume of data interaction between the network nodes included in the network node cluster and other network node clusters per unit time when the data processing load parameter of the network node cluster exceeds a preset load parameter threshold;

[0134] The correlation coefficient acquisition execution module is used to obtain the correlation coefficient of each network node with other network node clusters by using the data exchange data volume between the network nodes included in the network node cluster and other network node clusters per unit time.

[0135] The correlation coefficient between each network node and other network node clusters is obtained by the following formula:

[0136]

[0137] Among them, G ij represents the correlation coefficient between the i-th network node and the j-th network node cluster; C ij represents the average amount of data interaction per unit time between the i-th network node and the j-th network node cluster; C zj Indicates the average amount of total data interaction per unit time of the jth network node cluster; C0 represents the dynamic baseline interaction volume (the median of the total system data interaction volume). The above formula measures the degree of interaction dependence between node i and node cluster j, through 1-e -Cij / C0 Achieve nonlinear saturation response to data volume.

[0138] The working principle of the above technical solution is as follows: First, the calculated data processing load parameter of a network node cluster is compared with a preset load parameter threshold. If the threshold is exceeded, it indicates that the network node cluster is facing significant data processing pressure, and further analysis of its relationship with other network node clusters is required to find possible solutions. Once the network node cluster is determined to be overloaded, the amount of data exchanged per unit time between each network node in the cluster and other network node clusters is retrieved. This step aims to collect data that reflects the closeness of connections between network nodes, as the amount of data exchange, to a certain extent, reflects the degree of association between different network node clusters. The acquired amount of data exchange per unit time is used to calculate the correlation coefficient between each network node and other network node clusters. Using a specific algorithm (such as normalization based on the amount of data exchange), the data exchange volume is converted into a numerical value that intuitively reflects the degree of association, namely the correlation coefficient. The higher the coefficient, the closer the corresponding network node is to the other network node clusters.

[0139] The above technical solution achieves the following: Using the volume of data exchanged per unit time between a network node and other network node clusters as the basis for calculating the correlation coefficient accurately reflects the actual degree of connection between different network nodes. Data exchange volume is a direct and key indicator of inter-node relationships. The correlation coefficient calculated based on this volume can more accurately assess the connections between network node clusters, providing accurate data support for subsequent decision-making. Different network nodes interact with other network node clusters in varying ways. By analyzing the volume of data exchanged per unit time, it is possible to effectively distinguish the varying degrees of connection between each network node and its cluster. This avoids simplistic generalizations and allows for a more nuanced understanding of the complex relationships within the network structure, helping to more effectively address the issue of overloaded network node clusters. In the IoT environment, data exchange between network nodes changes dynamically with evolving business needs and the network environment. Calculating the correlation coefficient based on the volume of data exchanged per unit time allows for timely adaptation to these changes. As the data traffic distribution within the network changes, the correlation coefficient adjusts accordingly, enabling the system to make informed decisions based on the latest network status, demonstrating strong flexibility and adaptability. Different IoT application scenarios have distinct data interaction patterns and characteristics. By focusing on the volume of data interaction per unit time, the correlation coefficient can be determined to adapt to diverse business scenarios. Whether for data-intensive applications or those with high real-time requirements, the correlation coefficient can be accurately calculated based on actual data interaction, providing strong support for network optimization in different business scenarios. By calculating the correlation coefficient, nodes with high correlation to currently overloaded network node clusters can be quickly located from among numerous network nodes. This helps the system quickly identify potential resource allocation targets, reducing the time and cost of finding solutions and improving the system's efficiency in addressing issues such as network congestion. Accurate correlation coefficients can help the system develop more optimized resource allocation strategies. By routing target network nodes to the server belonging to the corresponding network node cluster based on the maximum correlation coefficient, network resources can be more rationally allocated, avoiding blind resource adjustments. This improves resource utilization efficiency across the IoT system and ensures efficient operation.

[0140] In one embodiment of the present invention, the target network node determination module includes:

[0141] A contribution intensity coefficient acquisition module is used to obtain the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which the network node is located by using the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located;

[0142] The contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is obtained by the following formula:

[0143]

[0144] Where S represents the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located; R represents the resource fluctuation load coefficient corresponding to each network node; R p represents the average value of resource fluctuation load coefficients corresponding to all network nodes in the network node cluster; δ represents the preset smoothing factor, which is used to prevent numerical instability when the denominator approaches zero, and its value range is 0.1-0.3; L represents the data processing load parameter of the network node cluster; L y Indicates the preset load parameter threshold;

[0145] a normalization processing module, configured to perform normalization processing on a contribution intensity coefficient of each network node to a data processing load parameter of the network node cluster in which it is located, so that the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is projected to an interval of (0, 1);

[0146] a ratio parameter acquisition module, configured to perform ratio processing on the data processing load parameter of the network node cluster and a preset load parameter threshold value, and obtain a ratio between the preset load parameter threshold value and the data processing load parameter of the network node cluster as a ratio parameter;

[0147] a load disturbance coefficient acquisition module, configured to acquire a load disturbance coefficient of each network node to the data processing load of the network node cluster in which it is located by using a contribution intensity coefficient and a ratio parameter of each network node to the data processing load parameter of the network node cluster in which it is located;

[0148] The load disturbance coefficient of each network node to the data processing load of the network node cluster in which it is located is obtained by the following formula:

[0149]

[0150] Among them, P r represents the load disturbance coefficient of each network node to the data processing load of the network node cluster where it is located; S represents the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster where it is located; G max Indicates the maximum value of the correlation coefficient between a network node and other network node clusters; G avg Represents the average value of the correlation coefficient between the network node and other network node clusters; B L Represents a ratio parameter.

[0151] The target network node screening module is used to select the network node corresponding to the maximum value of the load disturbance coefficient as the target network node.

[0152] The working principle of the above technical solution is as follows: First, the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located are comprehensively considered. A specific calculation method (such as weighted calculation) is used to derive the contribution intensity coefficient of each network node to the data processing load parameter of its network node cluster. This coefficient reflects the proportion or influence of each network node in the data processing load of the entire network node cluster. The calculated contribution intensity coefficient of each network node is normalized, and its numerical range is projected to the interval (0, 1). The purpose of normalization is to make the contribution intensity coefficients of different network nodes comparable and eliminate the influence caused by different original numerical ranges. The data processing load parameter of the network node cluster is compared with a preset load parameter threshold to obtain the ratio between the two, which is used as a ratio parameter. This ratio parameter can reflect the degree of deviation of the data processing load of the current network node cluster from the preset threshold. Combining the contribution intensity coefficient of each network node with the calculated ratio parameter, a specific algorithm (such as multiplication) is used to calculate the load disturbance coefficient of each network node to the data processing load of its network node cluster. The load perturbation coefficient comprehensively considers the node's own contribution and the load deviation of the entire node cluster. The load perturbation coefficients of all network nodes are compared, and the node with the largest load perturbation coefficient is identified as the target network node. Selecting the node with the largest load perturbation coefficient indicates that it has the greatest impact on the load of the network node cluster. Selecting it as the target node for subsequent processing (such as forwarding to a server belonging to another network node cluster) is expected to alleviate the current network node cluster's load pressure to the greatest extent.

[0153] The above technical solution achieves the following: By comprehensively considering multiple factors, including resource fluctuation load coefficients and network node cluster data processing load parameters, it accurately calculates each network node's contribution intensity coefficient and load disturbance coefficient to the load of its corresponding network node cluster. This allows the system to precisely locate the target network node with the greatest impact on the data processing load of the network node cluster, avoiding misselection and improving decision accuracy. The calculation of the load disturbance coefficient not only considers the node's own contribution but also incorporates the load deviation of the network node cluster, more accurately reflecting the actual impact of the network node on the load of the entire node cluster. This comprehensive approach ensures that the determined target network nodes better meet actual needs and provides a more reliable basis for subsequent load balancing and resource allocation. Selecting the network node with the highest load disturbance coefficient as the target network node can minimize the load pressure on the current network node cluster. Because this node has the greatest impact on the load, relocating it can quickly reduce the load of the network node cluster, improve system response speed, reduce data processing latency, and thus enhance the operational efficiency of the entire IoT system. Accurately determining the target network node allows for more targeted resource allocation. This avoids blind attempts to transfer different nodes, reduces ineffective operations during resource allocation, optimizes the resource allocation process, saves time and resources, and improves resource utilization efficiency. By rationally identifying the target network node and performing the transfer, the load distribution within the network node cluster is more evenly distributed, avoiding system instability caused by excessive load on individual nodes. This helps enhance the system's resistance to interference and improves its stability in the face of various network changes and load fluctuations. Effective load balancing and resource allocation can reduce the risk of overloading the network node cluster, reducing the possibility of system failures and service interruptions caused by overload, ensuring the continuous and stable operation of the IoT system and improving system reliability and availability.

[0154] On the other hand, the formula comprehensively considers the node's own contribution, its association characteristics with other node clusters, and the overall load deviation of the node cluster, comprehensively covering the key factors that affect the load disturbance of the network node to its node cluster. Compared with a single factor measurement, it can more accurately reflect the actual impact of the node on the load, provide a basis for accurately judging the role of the node on the network load, and improve the accuracy of load assessment. Ratio parameter B L The formula can be adaptively adjusted according to the different situations of network node cluster load relative to the threshold. Different load deviations from the threshold have different effects on the load disturbance coefficient calculation, making the calculation results more in line with the actual network operation status and further improving the accuracy. It can sensitively detect subtle differences in the degree of connectivity between a network node and other node clusters. Even if the inter-node connectivity changes slightly, this ratio can be detected and reflected in the load disturbance coefficient, making the system more sensitive to changes in node connectivity within the network structure. As the network operates, node contribution strength, connectivity, and cluster load dynamically change. The formula calculates the load disturbance coefficient based on these changes in real time, rapidly responding to changes in network status and promptly reflecting the dynamic changes in a node's impact on the load. Accurately calculating the load disturbance coefficient helps pinpoint nodes with the greatest impact on the load of network node clusters. The system can then rationally allocate resources, shift tasks from overloaded nodes, achieve load balancing, reduce system instability caused by uneven node load, and enhance system stability. By accurately assessing node load disturbances, the system can take proactive measures, such as adjusting task allocation and adding resources, before node or cluster loads reach dangerous levels, to prevent overload risks and ensure continued stable system operation.

[0155] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. An Internet of Things intelligent monitoring method, characterized in that: The Internet of Things intelligent monitoring method includes: Real-time monitoring of the data processing load parameters of the network node cluster corresponding to each server included in the Internet of Things; When the data processing load parameter of the network node cluster exceeds a preset load parameter threshold, obtaining the correlation coefficient between the network node included in the network node cluster and other network node clusters; Determining a target network node using a correlation coefficient between the network node and other network node clusters; The target network node is transferred to a server to which a network node cluster corresponding to the maximum value of the target network node association coefficient belongs.

2. The method for intelligent monitoring of the Internet of Things according to claim 1, characterized in that: Real-time monitoring of the data processing load parameters of the network node cluster corresponding to each server included in the Internet of Things, including: Real-time monitoring of the operating parameters of each network node included in the network node cluster corresponding to each server included in the Internet of Things; wherein the operating parameters of the network node include CPU utilization, memory utilization and cache hit rate; Real-time monitoring of the percentage change in CPU utilization and memory utilization corresponding to each percentage unit increase in cache hit rate; wherein the value range of the percentage unit corresponding to the cache hit rate is 0.1%-0.3%; Obtaining the resource fluctuation load coefficient corresponding to each network node using the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase of the cache hit rate; The data processing load parameter of the network node cluster is obtained by utilizing the resource fluctuation load coefficient corresponding to each network node included in the network node cluster.

3. The method for intelligent monitoring of the Internet of Things according to claim 2, characterized in that: The data processing load parameters of the network node cluster are obtained by using the resource fluctuation load coefficient corresponding to each network node included in the network node cluster, including: Retrieving the resource fluctuation load coefficient corresponding to each network node included in the network node cluster; Obtaining a standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster according to the resource fluctuation load coefficient corresponding to each network node included in the network node cluster; The data processing load parameter of the network node cluster is obtained by utilizing the resource fluctuation load coefficient corresponding to each network node included in the network node cluster and the standard deviation of the resource fluctuation load coefficient corresponding to the network node cluster.

4. The method for intelligent monitoring of the Internet of Things according to claim 1, characterized in that: When the data processing load parameter of the network node cluster exceeds a preset load parameter threshold, obtaining the correlation coefficient between the network node included in the network node cluster and other network node clusters includes: When the data processing load parameter of the network node cluster exceeds the preset load parameter threshold, the amount of data exchanged between the network nodes contained in the network node cluster and other network node clusters per unit time is retrieved. The correlation coefficient of each network node with other network node clusters is obtained by utilizing the data volume of data exchanged between the network nodes included in the network node cluster and other network node clusters per unit time.

5. The method for intelligent monitoring of the Internet of Things according to claim 1, characterized in that: Determining a target network node by using a correlation coefficient between the network node and other network node clusters includes: The contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which the network node is located is obtained by using the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located; Normalizing the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located so that the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is projected to the interval (0, 1); Performing ratio processing on the data processing load parameter of the network node cluster and a preset load parameter threshold value to obtain a ratio between the preset load parameter threshold value and the data processing load parameter of the network node cluster as a ratio parameter; Obtaining a load disturbance coefficient of the data processing load of the network node cluster to which each network node belongs by using the contribution intensity coefficient and the ratio parameter of each network node to the data processing load parameter of the network node cluster to which it belongs; The network node corresponding to the maximum value of the load disturbance coefficient is used as the target network node.

6. An Internet of Things intelligent monitoring system, characterized in that: The Internet of Things intelligent monitoring system includes: A monitoring module is used to monitor in real time the data processing load parameters of the network node cluster corresponding to each server included in the Internet of Things; a correlation coefficient acquisition module, configured to acquire correlation coefficients between the network nodes included in the network node cluster and other network node clusters when a data processing load parameter of the network node cluster exceeds a preset load parameter threshold; a target network node determination module, configured to determine a target network node by using a correlation coefficient between the network node and other network node clusters; The node switching module is used to switch the target network node to the server to which the network node cluster corresponding to the maximum value of the correlation coefficient belongs.

7. The IoT intelligent monitoring system according to claim 6, characterized in that: The monitoring module includes: A first parameter monitoring module is used to monitor in real time the operating parameters of each network node included in the network node cluster corresponding to each server included in the Internet of Things; wherein the operating parameters of the network node include CPU utilization, memory utilization and cache hit rate; The second parameter monitoring module is used to monitor in real time the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase in cache hit rate; wherein the value range of the percentage unit corresponding to the cache hit rate is 0.1%-0.3%; A resource fluctuation load coefficient acquisition module is used to obtain the resource fluctuation load coefficient corresponding to each network node by using the percentage change of CPU utilization and memory utilization corresponding to each percentage unit increase of the cache hit rate; The data processing load parameter acquisition module is used to acquire the data processing load parameter of the network node cluster by using the resource fluctuation load coefficient corresponding to each network node included in the network node cluster.

8. The IoT intelligent monitoring system according to claim 7, characterized in that: The data processing load parameter acquisition module includes: A resource fluctuation load coefficient retrieving module is used to retrieve the resource fluctuation load coefficient corresponding to each network node included in the network node cluster; A standard deviation acquisition module, configured to acquire a standard deviation of a resource fluctuation load coefficient corresponding to a network node cluster according to a resource fluctuation load coefficient corresponding to each network node included in the network node cluster; The parameter acquisition execution module is used to obtain the data processing load parameter of the network node cluster by using the resource fluctuation load coefficient corresponding to each network node included in the network node cluster and the resource fluctuation load coefficient standard deviation corresponding to the network node cluster.

9. The IoT intelligent monitoring system according to claim 6, characterized in that: The correlation coefficient acquisition module includes: An interactive data volume retrieving module, configured to retrieve the data volume of data interaction between the network nodes included in the network node cluster and other network node clusters per unit time when the data processing load parameter of the network node cluster exceeds a preset load parameter threshold; The correlation coefficient acquisition execution module is used to obtain the correlation coefficient of each network node with other network node clusters by using the data exchange data volume between the network nodes included in the network node cluster and other network node clusters per unit time.

10. The Internet of Things intelligent monitoring system according to claim 6, characterized in that: The target network node determination module includes: A contribution intensity coefficient acquisition module is used to obtain the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which the network node is located by using the resource fluctuation load coefficient corresponding to each network node and the data processing load parameter of the network node cluster in which the network node is located; a normalization processing module, configured to perform normalization processing on a contribution intensity coefficient of each network node to a data processing load parameter of the network node cluster in which it is located, so that the contribution intensity coefficient of each network node to the data processing load parameter of the network node cluster in which it is located is projected to an interval of (0, 1); a ratio parameter acquisition module, configured to perform ratio processing on the data processing load parameter of the network node cluster and a preset load parameter threshold value, and obtain a ratio between the preset load parameter threshold value and the data processing load parameter of the network node cluster as a ratio parameter; a load disturbance coefficient acquisition module, configured to acquire a load disturbance coefficient of each network node to the data processing load of the network node cluster in which it is located by using a contribution intensity coefficient and a ratio parameter of each network node to the data processing load parameter of the network node cluster in which it is located; The target network node screening module is used to select the network node corresponding to the maximum value of the load disturbance coefficient as the target network node.

Images