Real-time short message communication link verification and performance monitoring method

Abstract

The invention discloses a real-time short message communication link verification and performance monitoring method, which relates to the technical field of short message communication link verification and performance monitoring, and comprises the following steps of: receiving flow, namely a short message request, from a user or other network equipment, and sending the flow to a selected server by applying a preset load distribution algorithm; and acquiring server information after data traffic distribution, performing exception analysis processing on the acquired server information, and establishing a configuration efficiency classification and identification model. By configuring the efficiency classification and identification model, the data traffic configuration can be accurately divided into three levels of high efficiency, general and low efficiency, the real-time monitoring of the performance state and load distribution of the server is realized, the improper configuration is timely adjusted, and the system automatically optimizes the general and low efficiency configuration, so that the efficiency of the system is greatly improved. For example, resources are adjusted through an automatic expansion function, and the weight of the server is optimized to allocate more traffic to the server with excellent performance.

Description

Technical Field

[0001] The present invention relates to the technical field of SMS communication link verification and performance monitoring, and in particular to a real-time SMS communication link verification and performance monitoring method. Background Art

[0002] Real-time SMS communication link verification and performance monitoring refers to the use of technical means to detect and verify in real time whether the transmission path of the SMS service is operating normally, as well as to evaluate and monitor the performance of the SMS transmission process. This process includes ensuring that the SMS can be successfully sent to the target recipient within the scheduled time, while monitoring the key performance indicators of the SMS service, such as availability, latency, and success rate. Specifically, this type of monitoring usually requires the deployment of specialized software or systems to track each link of the SMS from sending to receiving in real time, so as to promptly discover possible problems in the link such as network delays and service interruptions. In addition, this type of monitoring can also help service providers analyze the overall efficiency and reliability of SMS services, thereby optimizing service quality and ensuring that user experience and business needs are met.

[0003] Real-time SMS communication link verification and performance monitoring use some signal transmission devices, including load balancers. The main function of load balancers is to distribute data traffic and requests entering the network to multiple servers or network devices to prevent any single device from being overloaded, thereby improving overall service reliability and efficiency. In the SMS communication system, the load balancer can ensure that the SMS service remains stable during high traffic periods, by optimizing resource allocation, reducing latency, and improving response speed. It also helps with failover and increases the system's fault tolerance. Such a configuration allows each server to operate at its optimal performance, thereby ensuring the continuity and quality of SMS transmission services.

[0004] The prior art has the following deficiencies:

[0005] When verifying and monitoring the performance of real-time SMS communication links, if the load balancer is improperly configured, it may not be able to effectively distribute network traffic, causing some servers to be overloaded while other servers are relatively idle. This uneven load distribution can cause servers with limited processing power to respond more slowly, leading to service delays or complete service interruptions. For example, if the delay in sending or receiving SMS is too long, it may affect the timeliness of emergency communications, such as the delivery of verification codes or emergency notifications.

[0006] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0007] The purpose of the present invention is to provide a real-time SMS communication link verification and performance monitoring method. By configuring an efficiency classification recognition model, the data traffic configuration can be accurately divided into three levels: high efficiency, general and low efficiency, thereby realizing real-time monitoring of server performance status and load distribution, and timely adjusting improper configurations. For general and inefficient configurations, the system automatically performs optimization processing, such as adjusting resources through automatic expansion functions, optimizing server weights to allocate more traffic to servers with superior performance. In addition, it also includes a dynamic monitoring mechanism to ensure that the optimized configuration can be upgraded to an efficient configuration in a timely manner. If the standard is not met, an early warning will be issued, thereby improving the stability, efficiency and response speed of the system to solve the problems in the above-mentioned background technology.

[0008] In order to achieve the above object, the present invention provides the following technical solution: a real-time SMS communication link verification and performance monitoring method, comprising the following steps:

[0009] Receives traffic, i.e., SMS requests, from users or other network devices and applies a preset load distribution algorithm to send the traffic to the selected server;

[0010] Obtain server information after data traffic allocation, perform abnormal analysis on the acquired server information, and establish a configuration efficiency classification and recognition model;

[0011] Based on the configuration efficiency classification recognition model, data traffic configuration is divided into efficient configuration, general configuration and inefficient configuration;

[0012] For inefficient configurations, automatic expansion functions are implemented through cloud services or modern data center technologies to automatically increase or decrease resources based on real-time load. For general configurations, weights are adjusted based on server performance and historical load data, and the weights of servers with excellent performance are increased to bear more traffic.

[0013] Monitor the process of optimizing inefficient configurations and general configurations. When the inefficient configurations and general configurations are transformed into efficient configurations after optimization, continue to configure data traffic. If they are not transformed into efficient configurations within a specified time, an early warning prompt will be issued.

[0014] Preferably, according to the selected algorithm, the specific steps of allocating traffic to the selected server are as follows:

[0015] Determine which load balancing strategy to use;

[0016] Configure the selected allocation algorithm;

[0017] Receive network requests from clients;

[0018] According to the preset algorithm, calculate which server each request should be sent to;

[0019] Based on the results of the algorithm, the request is forwarded to the selected server.

[0020] Preferably, the server information after data traffic allocation includes the backlog of pending requests in the server queue and the deviation between the actual allocated traffic and the predicted traffic. After performing abnormal analysis on the backlog of pending requests in the server queue, a queue depth index is generated. After performing abnormal analysis on the deviation between the actual allocated traffic and the predicted traffic, a traffic prediction deviation index is generated.

[0021] Preferably, after obtaining the queue depth index and traffic prediction deviation index generated by the server within T time after data traffic allocation, a configuration efficiency classification and identification model is established for the queue depth index and the traffic prediction deviation index to generate a configuration efficiency impact coefficient, and the configuration efficiency of the data traffic is intelligently evaluated through the configuration efficiency impact coefficient.

[0022] Preferably, the configuration efficiency impact coefficient generated by the server within T time after data traffic allocation is compared and analyzed with a preset first gradient reference threshold and a second gradient reference threshold, wherein the first gradient reference threshold is less than the second gradient reference threshold, and the result of the comparison and analysis is as follows:

[0023] If the configuration efficiency impact coefficient is less than the first gradient reference threshold, the data flow configuration within the T time is classified as a high-efficiency configuration;

[0024] If the configuration efficiency impact coefficient is greater than or equal to the first gradient reference threshold and less than the second gradient reference threshold, the data flow configuration within the T time is classified as a general configuration;

[0025] If the configuration efficiency impact coefficient is greater than or equal to the second gradient reference threshold, the data flow configuration within the T time is classified as an inefficient configuration.

[0026] Preferably, the automatic expansion function is implemented through cloud services or modern data center technology. The specific steps are as follows:

[0027] Choose a cloud platform that supports auto-scaling;

[0028] Create an Auto Scaling group in the AWS Management Console;

[0029] Set indicators that trigger expansion and define thresholds;

[0030] Specify the configuration to use when launching a new instance;

[0031] Use AWS CloudWatch for performance monitoring, set monitoring intervals and alarms, and configure alarms based on preset performance metric thresholds to automatically trigger scaling actions;

[0032] Deploy applications in an auto-scaling environment and ensure all configurations and dependencies are correct.

[0033] Preferably, a fuzzy logic control algorithm is used to adjust the server weights, and the specific steps are as follows:

[0034] Select key indicators reflecting server performance as input variables and the weight of each server as output variable;

[0035] Define fuzzy terms for each variable;

[0036] Formulate fuzzy rules based on system management experience and performance goals;

[0037] Use Mamdani as fuzzy reasoning method;

[0038] Reasoning is performed based on the defined fuzzy rules and the input real-time data to obtain the fuzzy weight output of each server;

[0039] Select defuzzification techniques to convert fuzzy outputs into precise weighted values ​​for practical applications.

[0040] According to the defuzzification results, the weight of each server in the load balancer is adjusted.

[0041] Preferably, when optimizing inefficient configurations and general configurations, the configuration efficiency impact coefficient generated within each T time is obtained in real time, and the configuration efficiency impact coefficient is compared and analyzed with the first gradient reference threshold. If there is a configuration efficiency impact coefficient less than the first gradient reference threshold within the specified N T times, the data flow configuration continues. If there is no configuration efficiency impact coefficient less than the first gradient reference threshold within the specified N T times, an early warning prompt is issued to notify relevant maintenance personnel to inspect the data flow configuration of the load balancer in advance.

[0042] Preferably, the logic for obtaining the queue depth index is as follows:

[0043] Get the real-time queue depth of the server at time t after data traffic allocation, that is, the number of requests to be processed at that time point, and represent the real-time queue depth with function Q(t);

[0044] Calculate the standard deviation of the queue depth using the expression: σ Q The standard deviation of the queue depth, which indicates the degree of fluctuation of the queue depth around the average value;

[0045] Calculate the coefficient of variation of the queue depth. The expression is: In the formula, CV Q The coefficient of variation of the queue depth is used to measure the relative fluctuation of the queue depth;

[0046] Calculate the queue depth index. The calculation expression is:

[0047] represents the maximum difference in queue depth over the mean value within the time window T, represents the logarithmic transformation of the maximum excess, adding a nonlinear measure to highlight abnormally high queue peaks, Q DI Indicates the queue depth index.

[0048] Preferably, the logic for obtaining the traffic prediction deviation index is as follows:

[0049] Obtain the actual traffic of the server at time t after data traffic allocation, and represent the actual traffic in time series using function A(t);

[0050] Establish an ARIMA model and fit the model. The specific formula of the model is:

[0051] Where q is the order of the moving average model part, which specifies the number of lagged forecast error terms included in the model, p is the order of the autoregressive model part, which specifies the number of historical data points included in the model, c is the constant term in the model, and φ i is the coefficient of the autoregressive part, which is associated with the value of the variable at time (ti), θ j is the coefficient of the moving average part, which is associated with the prediction error at time (tj), and ∈(t) is the prediction error term at time t;

[0052] Use the fitted ARIMA model to predict the predicted flow P(t) within the time window T, and calculate the deviation D(t) between the actual flow and the predicted flow. The calculation expression is: D(t) = A(t) - P(t);

[0053] The flow prediction deviation index is calculated based on the deviation D(t) between the actual flow and the predicted flow. The calculation expression is: Where TP DI Represents the traffic prediction deviation index.

[0054] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0055] The present invention can accurately divide data traffic configuration into three levels: efficient configuration, general configuration and inefficient configuration by establishing a configuration efficiency classification identification model. This classification enables system administrators to monitor the performance status and load distribution of each server in real time, and promptly discover and adjust improper configuration situations. When the configuration efficiency of a server is in the general configuration and inefficient configuration, the system will automatically perform optimization processing to ensure that all servers can process traffic evenly and effectively, thereby avoiding overload and waste of resources, and improving the stability and response speed of the entire system.

[0056] The present invention proposes corresponding optimization measures. For inefficient configurations, the automatic expansion function is implemented to automatically adjust resource allocation to ensure service continuity and efficiency. For general configurations, the server weight is adjusted to optimize traffic distribution so that servers with better performance can handle more traffic. In addition, a dynamic monitoring mechanism is included to ensure that general or inefficient configurations can be upgraded to efficient configurations in a timely manner after optimization. If the optimized configuration fails to reach the standard of efficient configuration within the specified time, an early warning prompt will be issued to promptly notify maintenance personnel to intervene, which not only improves the stability and efficiency of the communication link, but also greatly enhances the service provider's control ability and response speed over network conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0058] Figure 1 The present invention is a method flow chart of a real-time SMS communication link verification and performance monitoring method. DETAILED DESCRIPTION

[0059] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0060] The present invention provides Figure 1 A real-time SMS communication link verification and performance monitoring method is shown, comprising the following steps:

[0061] Receives traffic, i.e., SMS requests, from users or other network devices and applies a preset load distribution algorithm to send the traffic to the selected server;

[0062] A load balancer decides which servers to forward received traffic to based on a built-in or preset algorithm. This decision is based on an algorithm designed to optimize resource usage and response time.

[0063] There are many load distribution algorithms configured inside the load balancer, such as:

[0064] Round Robin: Allocate requests to each server in the server list in order.

[0065] Least Connections: Prioritize sending requests to the server with the least number of active connections.

[0066] IP Hashing: Generates a hash value based on the IP address of the request source and directs traffic to a specific server based on this hash value.

[0067] The purpose of applying these algorithms is to ensure that server resources are used reasonably, avoiding overloading of some servers while others are idle. This helps improve the overall service response speed and reliability, ensuring that user requests are processed quickly and in a balanced manner.

[0068] According to the selected algorithm, the specific steps to distribute traffic to the selected servers are as follows:

[0069] Determine which load balancing strategy to use;

[0070] Common strategies include Round Robin, Least Connections, IP Hashing, Resource Utilization, etc. Choose the most suitable allocation algorithm based on application requirements and server environment. For example, if the server performance is similar, you may choose Round Robin; if the server performance is uneven, you may choose the least connection or resource utilization strategy.

[0071] Configure the selected allocation algorithm;

[0072] This usually involves setting specific parameters of the algorithm, such as weights (in round-robin), connection thresholds (in least connections), etc.

[0073] Receive network requests from clients;

[0074] These requests can be HTTP requests, database requests, or any other type of network traffic.

[0075] According to the preset algorithm, calculate which server each request should be sent to;

[0076] For example, the round-robin algorithm selects servers in order, and the least-connected algorithm finds the server with the fewest current connections.

[0077] Based on the results of the algorithm, the request is forwarded to the selected server;

[0078] During the forwarding process, it may modify the header information of the data packet to ensure that the data is correctly sent to the destination server.

[0079] Obtain server information after data traffic allocation, perform abnormal analysis on the acquired server information, and establish a configuration efficiency classification and recognition model;

[0080] The server information after data traffic allocation includes the backlog of pending requests in the server queue and the deviation between the actual allocated traffic and the predicted traffic (predicted by the load balancer). After performing abnormal analysis on the backlog of pending requests in the server queue, a queue depth index is generated. After performing abnormal analysis on the deviation between the actual allocated traffic and the predicted traffic, a traffic prediction deviation index is generated.

[0081] A backlog of pending requests in the server queues indicates that the load balancer may be misconfigured for data traffic. In an ideal load balancing setup, the load balancer should be able to evenly distribute incoming requests based on the current load and processing capacity of each server to avoid overloading any single server. If you observe a persistent backlog of requests in some servers' queues while other servers are relatively idle, this is usually an indication that the load balancer's allocation algorithm may not be executing correctly or that its configuration is not appropriate for the current network and server conditions. This uneven load distribution can lead to increased processing delays, affect user experience, and increase the risk of server overload, thereby affecting the performance and reliability of the entire system. Therefore, this backlog is an important warning sign that the load balancer configuration needs to be adjusted or its allocation strategy needs to be improved.

[0082] The logic for obtaining the queue depth index is as follows:

[0083] Get the real-time queue depth of the server at time t after data traffic allocation, that is, the number of requests to be processed at that time point, and represent the real-time queue depth with function Q(t);

[0084] New Relic provides comprehensive performance monitoring of servers and applications, allowing real-time viewing of the status and depth of application queues.

[0085] Calculate the standard deviation of the queue depth using the expression: σ Q The standard deviation of the queue depth, which indicates the degree of fluctuation of the queue depth around the average value;

[0086] Calculate the coefficient of variation of the queue depth. The expression is: In the formula, CV Q The coefficient of variation of the queue depth is used to measure the relative fluctuation of the queue depth;

[0087] Calculate the queue depth index. The calculation expression is:

[0088] represents the maximum difference in queue depth over the mean value within the time window T, represents the logarithmic transformation of the maximum excess, adding a nonlinear measure to highlight abnormally high queue peaks, Q DI Indicates the queue depth index.

[0089] From the calculation expression of the queue depth variation coefficient, it can be seen that the larger the performance value of the queue depth index generated by the server within T time after the data traffic is distributed, the greater the risk of improper data traffic configuration of the load balancer. Conversely, the smaller the risk of improper data traffic configuration of the load balancer.

[0090] Deviations between actual allocated traffic and predicted traffic indicate that the load balancer has a hidden danger of improper data traffic configuration. The main function of a load balancer is to balance and optimize data traffic between servers based on predetermined algorithms and prediction models to ensure that each server can effectively handle its assigned load. If a significant deviation between actual and predicted traffic is observed, this may be due to a variety of reasons, including inaccurate configuration of the load balancer's algorithm, insufficient precision of the prediction model, or failure to take into account all influencing factors (such as sudden traffic peaks, server performance changes, etc.). This deviation may cause some servers to be overloaded and others to be idle, affecting the performance and efficiency of the entire system, and may even cause service interruptions or delays.

[0091] The logic for obtaining the traffic prediction deviation index is as follows:

[0092] Obtain the actual traffic of the server at time t after data traffic allocation, and represent the actual traffic in time series using function A(t);

[0093] SNMP is a widely used network management protocol that collects and organizes information about network devices such as routers, switches, and servers, and modifies the behavior of devices to manage network performance. By setting up an SNMP agent, you can collect server traffic data in real time, and then monitor and analyze it through a network management system.

[0094] Establish an ARIMA model and fit the model. The specific formula of the model is:

[0095] Where q is the order of the moving average model part, which specifies the number of lagged forecast error terms included in the model, p is the order of the autoregressive model part, which specifies the number of historical data points (or lags) included in the model, c is the constant term in the model, and φ i is the coefficient of the autoregressive part, which is associated with the value of the variable at time (ti), θ j is the coefficient of the moving average part, which is associated with the prediction error at time (tj), and ∈(t) is the prediction error term at time t;

[0096] Use the fitted ARIMA model to predict the predicted flow P(t) within the time window T, and calculate the deviation D(t) between the actual flow and the predicted flow. The calculation expression is: D(t) = A(t) - P(t);

[0097] The flow prediction deviation index is calculated based on the deviation D(t) between the actual flow and the predicted flow. The calculation expression is: Where TP DI Represents the traffic prediction deviation index.

[0098] From the calculation expression of the traffic prediction deviation index, it can be seen that the larger the performance value of the traffic prediction deviation index generated by the server within T time after the data traffic is allocated, the greater the risk of improper data traffic configuration of the load balancer. Conversely, the smaller the risk of improper data traffic configuration of the load balancer.

[0099] The queue depth index Q generated by the server within T time after obtaining data traffic allocation DI and flow prediction deviation index TP DI Then, the queue depth index Q DI and flow prediction deviation index TP DI Establish a classification and identification model for configuration efficiency and generate the configuration efficiency impact coefficient Con μ , through the configuration efficiency influence coefficient Con μ Intelligently evaluate the configuration efficiency of data traffic.

[0100] The configuration efficiency classification recognition model is not specifically limited here, and can achieve the queue depth index Q DI and flow prediction deviation index TP DI Conduct comprehensive analysis and generate the configuration efficiency impact coefficient Con μ In order to realize the technical solution of the present invention, the present invention provides a specific implementation method;

[0101] Configuration efficiency influence coefficient Con μ The resulting calculation formula is:

[0102] Conμ =(d1*Q DI +d2*TP DI )*(d1+d2), where d1 and d2 are queue depth index Q DI and flow prediction deviation index TP DI The preset proportional coefficient, and d1 and d2 are both greater than 0.

[0103] From the calculation formula, we can see that the larger the queue depth index generated by the server within T time after data traffic allocation, the larger the traffic prediction deviation index, that is, the configuration efficiency influence coefficient Con generated by the server within T time after data traffic allocation μ The larger the performance value, the greater the risk of improper data traffic configuration of the load balancer. Conversely, the smaller the risk of improper data traffic configuration of the load balancer.

[0104] Based on the configuration efficiency classification recognition model, data traffic configuration is divided into efficient configuration, general configuration and inefficient configuration;

[0105] The configuration efficiency impact coefficient generated by the server within T time after data traffic allocation is compared and analyzed with the preset first gradient reference threshold and the second gradient reference threshold, wherein the first gradient reference threshold is less than the second gradient reference threshold. The results of the comparison and analysis are as follows:

[0106] If the configuration efficiency impact coefficient is less than the first gradient reference threshold, it indicates that the load balancer has a high configuration efficiency for data traffic, and the data traffic configuration within the T time is classified as an efficient configuration;

[0107] If the configuration efficiency impact coefficient is greater than or equal to the first gradient reference threshold and less than the second gradient reference threshold, it indicates that the configuration efficiency of the load balancer for the data traffic is general, and the data traffic configuration within the T time is classified as general configuration;

[0108] If the configuration efficiency impact coefficient is greater than or equal to the second gradient reference threshold, indicating that the load balancer has low configuration efficiency for data traffic, the data traffic configuration within the T time is classified as an inefficient configuration.

[0109] For inefficient configurations, auto-scaling capabilities can be implemented through cloud services or modern data center technologies to automatically increase or decrease resources based on real-time load.

[0110] To achieve automatic expansion through cloud services or modern data center technology, the specific steps are as follows:

[0111] Choose a cloud platform that supports auto-scaling;

[0112] Automatic scaling supports cloud platforms such as Amazon Web Services, Google Cloud Platform, or Microsoft Azure, while setting up cloud service accounts and configuring necessary access permissions and security settings.

[0113] Create an Auto Scaling group in the AWS Management Console;

[0114] This includes selecting the appropriate instance type, configuring the AMI (Amazon Machine Image), defining network and subnet settings, etc.

[0115] Set the indicators that trigger the expansion (such as CPU utilization, network I / O, disk I / O, etc.) and define the thresholds;

[0116] For example, you can set up a new instance to be added automatically when the CPU utilization exceeds 70%.

[0117] Specify the configuration to use when launching a new instance;

[0118] The configuration used includes instance type, AMI, associated security groups, IAM roles, etc., which is similar to the launch configuration, but provides more customization options and version control.

[0119] Use AWS CloudWatch for performance monitoring, set monitoring intervals and alarms, and configure alarms based on preset performance metric thresholds to automatically trigger scaling actions;

[0120] AWS CloudWatch is a monitoring service for tracking AWS cloud resources and applications running on AWS. It can collect and track metrics, collect and monitor log files, set alarms, and automatically respond to changes in the AWS environment;

[0121] Set up monitoring intervals and alarms in AWS CloudWatch, and configure alarms to automatically trigger scaling actions based on preset performance metric thresholds:

[0122] First, it involves selecting the metrics to be monitored in CloudWatch, such as the CPU utilization or memory usage of EC2 instances; then, setting the monitoring interval to determine how often CloudWatch collects data for these metrics, usually every minute or every five minutes; then, creating alarms in CloudWatch to set thresholds for these metrics, such as CPU utilization exceeding 80%. When the actual monitoring data exceeds these thresholds, the CloudWatch alarm is triggered. These alarms can be configured to send notifications to Amazon SNS and trigger policies associated with AWS Auto Scaling to automatically adjust the number of EC2 instances, increase to cope with high loads, or reduce to save costs, thereby achieving automatic expansion and reduction of resources. This entire process implements resource management based on actual application needs, optimizing performance and cost efficiency.

[0123] Deploy the application in an auto-scaling environment and ensure all configurations and dependencies are correct;

[0124] Use CloudWatch and other management tools to continuously monitor system performance, ensure smooth application operation, and adjust expansion strategies in a timely manner to cope with possible changes.

[0125] For general configurations, adjust the weights based on the server's performance and historical load data, increase the weights of servers with excellent performance, and thus handle more traffic;

[0126] Use fuzzy logic control algorithm to adjust server weights so that servers with good performance can handle more traffic. The specific steps are as follows:

[0127] Select key indicators reflecting server performance as input variables and the weight of each server as output variable;

[0128] Input variables such as CPU utilization, memory usage, response time, etc.

[0129] Define fuzzy terms for each variable (e.g., low, medium, high);

[0130] For example, CPU utilization may be classified as low (0-40%), medium (30%-70%), high (60%-100%).

[0131] Formulate fuzzy rules based on system management experience and performance goals;

[0132] For example, if CPU utilization is low and memory usage is low, the weight is high. The rule can use logical operators such as AND and OR to connect multiple conditions.

[0133] Use Mamdani as fuzzy reasoning method;

[0134] The most common ones are Mamdani and Takagi-Sugeno. The Mamdani method is the most commonly used because it is intuitive and easy to understand.

[0135] Reasoning is performed based on the defined fuzzy rules and the input real-time data to obtain the fuzzy weight output of each server;

[0136] Select defuzzification technology to convert fuzzy output into accurate weighted values ​​for practical application;

[0137] Commonly used defuzzification techniques include the centroid method (most commonly used), the bilateral maximum method, the median maximum method, etc.

[0138] According to the defuzzification results, the weight of each server in the load balancer is adjusted.

[0139] Monitor the process of optimizing inefficient configuration and general configuration (optimizing data flow configuration). When the inefficient configuration and general configuration are transformed into efficient configuration after optimization, continue to configure data flow. If they are not transformed into efficient configuration within the specified time, issue an early warning.

[0140] When optimizing inefficient configurations and general configurations, obtain the configuration efficiency impact coefficient generated within each T time in real time, and compare and analyze the configuration efficiency impact coefficient with the first gradient reference threshold. If there is a configuration efficiency impact coefficient less than the first gradient reference threshold within the specified N T times, continue to configure the data flow. If there is no configuration efficiency impact coefficient less than the first gradient reference threshold within the specified N T times, an early warning prompt is issued to notify relevant maintenance personnel to inspect the data flow configuration of the load balancer in advance.

[0141] The present invention can accurately divide data traffic configuration into three levels: efficient configuration, general configuration and inefficient configuration by establishing a configuration efficiency classification identification model. This classification enables system administrators to monitor the performance status and load distribution of each server in real time, and promptly discover and adjust improper configuration situations. When the configuration efficiency of a server is in the general configuration and inefficient configuration, the system will automatically perform optimization processing to ensure that all servers can process traffic evenly and effectively, thereby avoiding overload and waste of resources, and improving the stability and response speed of the entire system.

[0142] The present invention proposes corresponding optimization measures. For inefficient configurations, the automatic expansion function is implemented to automatically adjust resource allocation to ensure service continuity and efficiency. For general configurations, the server weight is adjusted to optimize traffic distribution so that servers with better performance can handle more traffic. In addition, a dynamic monitoring mechanism is included to ensure that general or inefficient configurations can be upgraded to efficient configurations in a timely manner after optimization. If the optimized configuration fails to reach the standard of efficient configuration within the specified time, an early warning prompt will be issued to promptly notify maintenance personnel to intervene, which not only improves the stability and efficiency of the communication link, but also greatly enhances the service provider's control ability and response speed over network conditions.

[0143] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0144] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0145] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A real-time SMS communication link verification and performance monitoring method, characterized in that: The following steps are involved: Receives traffic, i.e., SMS requests, from users or other network devices and applies a preset load distribution algorithm to send the traffic to the selected server; Obtain server information after data traffic allocation, perform abnormal analysis on the acquired server information, and establish a configuration efficiency classification and recognition model; Based on the configuration efficiency classification recognition model, data traffic configuration is divided into efficient configuration, general configuration and inefficient configuration; For inefficient configurations, automatic expansion functions are implemented through cloud services or modern data center technologies to automatically increase or decrease resources based on real-time load. For general configurations, weights are adjusted based on server performance and historical load data, and the weights of servers with excellent performance are increased to bear more traffic. Monitor the process of optimizing inefficient configurations and general configurations. When the inefficient configurations and general configurations are transformed into efficient configurations after optimization, continue to configure data traffic. If they are not transformed into efficient configurations within a specified time, an early warning prompt will be issued.

2. A real-time SMS communication link verification and performance monitoring method according to claim 1, characterized in that: According to the selected algorithm, the specific steps to distribute traffic to the selected servers are as follows: Determine which load balancing strategy to use; Configure the selected allocation algorithm; Receive network requests from clients; According to the preset algorithm, calculate which server each request should be sent to; Based on the results of the algorithm, the request is forwarded to the selected server.

3. A real-time SMS communication link verification and performance monitoring method according to claim 1, characterized in that: The server information after data traffic allocation includes the backlog of pending requests in the server queue and the deviation between the actual allocated traffic and the predicted traffic. After performing abnormal analysis on the backlog of pending requests in the server queue, a queue depth index is generated. After performing abnormal analysis on the deviation between the actual allocated traffic and the predicted traffic, a traffic prediction deviation index is generated.

4. A real-time SMS communication link verification and performance monitoring method according to claim 3, characterized in that: After obtaining the queue depth index and traffic prediction deviation index generated by the server within T time after data traffic allocation, a configuration efficiency classification and identification model is established based on the queue depth index and traffic prediction deviation index to generate a configuration efficiency impact coefficient. The configuration efficiency of data traffic is intelligently evaluated through the configuration efficiency impact coefficient.

5. A real-time SMS communication link verification and performance monitoring method according to claim 4, characterized in that: The configuration efficiency impact coefficient generated by the server within T time after data traffic allocation is compared and analyzed with the preset first gradient reference threshold and the second gradient reference threshold, wherein the first gradient reference threshold is less than the second gradient reference threshold. The results of the comparison and analysis are as follows: If the configuration efficiency impact coefficient is less than the first gradient reference threshold, the data flow configuration within the T time is classified as a high-efficiency configuration; If the configuration efficiency impact coefficient is greater than or equal to the first gradient reference threshold and less than the second gradient reference threshold, the data flow configuration within the T time is classified as a general configuration; If the configuration efficiency impact coefficient is greater than or equal to the second gradient reference threshold, the data flow configuration within the T time is classified as an inefficient configuration.

6. A real-time SMS communication link verification and performance monitoring method according to claim 1, characterized in that: To achieve automatic expansion through cloud services or modern data center technology, the specific steps are as follows: Choose a cloud platform that supports auto-scaling; Create an Auto Scaling group in the AWS Management Console; Set indicators that trigger expansion and define thresholds; Specify the configuration to use when launching new instances; Use AWS CloudWatch for performance monitoring, set monitoring intervals and alarms, and configure alarms based on preset performance metric thresholds to automatically trigger scaling actions; Deploy applications in an auto-scaling environment and ensure all configurations and dependencies are correct.

7. A real-time SMS communication link verification and performance monitoring method according to claim 1, characterized in that: Use fuzzy logic control algorithm to adjust server weights. The specific steps are as follows: Select key indicators reflecting server performance as input variables and the weight of each server as output variable; Define fuzzy terms for each variable; Formulate fuzzy rules based on system management experience and performance goals; Use Mamdani as fuzzy reasoning method; Reasoning is performed based on the defined fuzzy rules and the input real-time data to obtain the fuzzy weight output of each server; Select defuzzification technology to convert fuzzy output into accurate weighted values ​​for practical application; According to the defuzzification results, the weight of each server in the load balancer is adjusted.

8. A real-time SMS communication link verification and performance monitoring method according to claim 5, characterized in that: When optimizing inefficient configurations and general configurations, obtain the configuration efficiency impact coefficient generated within each T time in real time, and compare and analyze the configuration efficiency impact coefficient with the first gradient reference threshold. If there is a configuration efficiency impact coefficient less than the first gradient reference threshold within the specified N T times, continue to configure the data flow. If there is no configuration efficiency impact coefficient less than the first gradient reference threshold within the specified N T times, an early warning prompt is issued to notify relevant maintenance personnel to inspect the data flow configuration of the load balancer in advance.

9. A real-time SMS communication link verification and performance monitoring method according to claim 3, characterized in that: The logic for obtaining the queue depth index is as follows: Get the real-time queue depth of the server at time t after data traffic allocation, that is, the number of requests to be processed at that time point, and represent the real-time queue depth with function Q(t); Calculate the standard deviation of the queue depth using the expression: σ Q The standard deviation of the queue depth, which indicates the degree of fluctuation of the queue depth around the average value; Calculate the coefficient of variation of the queue depth. The expression for calculation is: In the formula, CV Q The coefficient of variation of the queue depth is used to measure the relative fluctuation of the queue depth; Calculate the queue depth index. The calculation expression is: represents the maximum difference in queue depth over the mean value within the time window T, represents the logarithmic transformation of the maximum excess, adding a nonlinear measure to highlight abnormally high queue peaks, Q DI Indicates the queue depth index.

10. A real-time SMS communication link verification and performance monitoring method according to claim 3, characterized in that: The logic for obtaining the traffic prediction deviation index is as follows: Obtain the actual traffic of the server at time t after data traffic allocation, and represent the actual traffic in time series using function A(t); Establish an ARIMA model and fit the model. The specific formula of the model is: Where q is the order of the moving average model part, which specifies the number of lagged forecast error terms included in the model, p is the order of the autoregressive model part, which specifies the number of historical data points included in the model, c is the constant term in the model, and φ i is the coefficient of the autoregressive part, which is associated with the value of the variable at time (ti), θ j is the coefficient of the moving average part, which is associated with the prediction error at time (tj), and ∈(t) is the prediction error term at time t; Use the fitted ARIMA model to predict the predicted flow P(t) within the time window T, and calculate the deviation D(t) between the actual flow and the predicted flow. The calculation expression is: D(t) = A(t) - P(t); The flow prediction deviation index is calculated based on the deviation D(t) between the actual flow and the predicted flow. The calculation expression is: Where TP DI Represents the traffic prediction deviation index.

Images