A monitoring, analyzing and controlling method for a paperless conference system

Abstract

This invention belongs to the field of high-definition video communication technology. It provides a monitoring, analysis, and control method for a paperless conferencing system, comprising: Step 1: Performing time delay analysis on the speaker switching process of a video conference to identify abnormal switching processes; Step 2: Analyzing the TCP connection establishment time based on the video conference logs to determine if a handshake delay exists; Step 3: If a handshake delay exists, matching the timestamps of the abnormal switching process with the handshake delay to determine if the abnormal switching process is caused by the handshake delay; Step 4: If so, analyzing the bandwidth data corresponding to the handshake delay during the abnormal switching process to determine the bandwidth threshold corresponding to the handshake delay; Step 5: Monitoring and analyzing the bandwidth change trend during the video conference switching process, and predicting the timing of the handshake delay occurrence based on the bandwidth threshold corresponding to the handshake delay, thereby ensuring the stability and smoothness of the video conference and improving the user experience.

Description

Technical Field

[0001] This invention belongs to the field of high-definition video communication technology, specifically a monitoring, analysis and control method for a paperless conferencing system. Background Technology

[0002] In the context of the rapid development of digital office and remote collaboration, video conferencing and digital paperless meetings have become important tools for enterprises in areas such as online communication, remote education, and online healthcare. However, the complexity and uncertainty of the network environment often lead to problems such as audio and video stuttering and black screens when switching speakers in video conferences, which seriously affect user experience and collaboration efficiency. Traditional conferencing systems have limitations in network anomaly detection and optimization, such as a lack of refined analysis of network latency during switching, difficulty in accurately locating the root cause of the problem, and inability to predict and proactively adjust in advance, resulting in frequent problems such as handshake delays and insufficient bandwidth. Therefore, improving user experience and ensuring the stability and smoothness of meetings have become urgent issues to be addressed.

[0003] Therefore, the present invention provides a monitoring, analysis and control method for a paperless meeting system. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0005] The technical solution adopted by this invention to solve its technical problem is: A monitoring, analysis, and control method for a paperless meeting system includes: Step 1: Perform time delay analysis on the speaker switching process in the video conference to identify abnormal switching processes; Step 2: Analyze the TCP connection establishment time based on the video conference logs to determine if there is a handshake delay; Step 3: If a handshake delay exists, timestamp the abnormal handover process and the handshake delay to determine whether the abnormal handshake process was caused by the handshake delay. Step 4: If so, analyze the bandwidth data corresponding to the handshake delay during the abnormal handover process to determine the bandwidth threshold corresponding to the handshake delay; Step 5: Monitor and analyze the bandwidth change trend during the video conference switching process, and predict the timing of handshake delay by combining the bandwidth threshold corresponding to the handshake delay.

[0006] Furthermore, the method for identifying abnormal switching processes is as follows: Obtain all round-trip times during the handover process, calculate the mean and standard deviation of the round-trip times, and calculate the fluctuation coefficient of the round-trip time by proportionally dividing the standard deviation by the mean. The difference between each round-trip time in the switching process and the round-trip time baseline is calculated, and the calculation results are integrated according to the time sequence to obtain the change sequence of round-trip time. For each element in the change sequence, if the element does not meet the requirements, it is marked as an abnormal element; otherwise, it is marked as a normal element. The number of intervals between anomalous elements in the round-trip time change sequence is counted, and the ratio of this number to the upper limit of the number of intervals is calculated to obtain the anomalous interval ratio. The maximum number of intervals is the total number of abnormal elements minus 1. The total number of normal elements that are between abnormal elements is counted, and the ratio of this number to the number of intervals between abnormal elements is calculated to obtain the mean of abnormal intervals. The ratio of the mean of abnormal intervals to the total number of elements in the changed sequence is then calculated to obtain the ratio of the number of abnormal intervals. The abnormal continuity value of round-trip time is obtained by summing the ratio of the number of abnormal intervals and the ratio of the number of abnormal intervals. If the fluctuation coefficient of the round-trip time is greater than the preset fluctuation threshold, and the abnormal continuous value is less than the preset continuous threshold, the handover process will be marked as an abnormal handover process.

[0007] Furthermore, the process of determining whether a handshake delay exists is as follows: Obtain the key timestamp data of the three-way handshake in the TCP connection establishment process, and get the time of the three handshakes. The handshake duration is the sum of the three handshake times. The handshake time is compared with the standard handshake time value. If the handshake time is greater than or equal to the standard time value, then there is a handshake delay.

[0008] Furthermore, the method for determining whether the abnormal handover process is caused by handshake delay is as follows: Obtain key timestamps of the abnormal handover process, including the handover trigger time, the actual handover completion time, and key timestamps of handshake delay, including the sending and receiving timestamps of the three-way handshake; The abnormal handover time window and handshake delay time window are calculated based on the key timestamp; Calculate the overlap duration of the abnormal handover time window and the handshake delay time window, and calculate the window overlap degree by proportionally calculating the abnormal handover window duration. If the window overlap is greater than or equal to the overlap threshold, and the handshake completion time is before the handover completion time, then the abnormal handover process is caused by handshake delay.

[0009] Furthermore, the abnormal handover time window is from the handover trigger time to the actual handover completion time, and the handshake delay time window is from the first handshake sending time to the third handshake receiving time.

[0010] Furthermore, the bandwidth threshold corresponding to the handshake delay is obtained as follows: Obtain the bandwidth corresponding to the handshake delay during the historical handover process, generate a bandwidth sequence, calculate the mean and standard deviation of the bandwidth sequence, and calculate the bandwidth fluctuation amplitude by proportionally dividing the standard deviation of the bandwidth sequence by the mean. The bandwidth fluctuation amplitude is compared with the bandwidth threshold. If it is less than or equal to the bandwidth threshold, the mean of the bandwidth sequence is used as the bandwidth threshold; otherwise, the minimum value in the bandwidth sequence is used as the bandwidth threshold.

[0011] Furthermore, the method for predicting the timing of the handshake delay is as follows: The critical bandwidth sequence is obtained by comparing the current bandwidth with the set bandwidth threshold, and the correlation coefficient of the critical bandwidth sequence is calculated based on the Pearson correlation coefficient method. The correlation coefficient of the critical bandwidth sequence is compared with the preset correlation coefficient. If the correlation coefficient meets the requirements, the critical bandwidth sequence is linear; otherwise, the critical bandwidth sequence is nonlinear. The timing of handshake delay can be predicted using linear and nonlinear critical bandwidth sequences.

[0012] Furthermore, the critical bandwidth sequence is obtained as follows: The bandwidth of the current video conference is monitored in real time, a bandwidth threshold is set, and the current bandwidth is compared with the bandwidth threshold. If the current bandwidth exceeds the bandwidth threshold, the difference between the current bandwidth and the bandwidth threshold is calculated to obtain the approximate value. The approach value is compared with the approach threshold. If the approach value is less than or equal to the approach threshold, the time point when the current bandwidth reaches the bandwidth threshold is marked as the critical time point. The time period between the critical time point and the current time point is marked as the over-limit period. The bandwidth threshold values ​​within the over-limit period are integrated in time sequence to obtain the critical bandwidth sequence.

[0013] Furthermore, the method of predicting the handshake delay occurrence time using a linear critical bandwidth sequence is as follows: If the critical bandwidth is linear, the least squares method is used to linearly fit the critical bandwidth sequence to obtain the handshake delay prediction model. The bandwidth threshold is used as input, and the handshake delay prediction model outputs the predicted time point when the bandwidth reaches the bandwidth threshold, which is the time when the handshake delay occurs.

[0014] Furthermore, the method of predicting the handshake delay occurrence time using a nonlinear critical bandwidth sequence is as follows: Random forests are used to predict the time when bandwidth reaches the bandwidth threshold, and a feature matrix is ​​constructed as input variables, including real-time bandwidth and critical bandwidth sequences. The input variables are normalized to the [0,1] interval. Each decision tree in the random forest performs forward inference on the input variables to obtain the prediction time of a single decision tree. The final prediction time is the average of the predictions of all decision trees, which is the time when the handshake delay occurs.

[0015] The beneficial effects of this invention are as follows: By analyzing the time delay of the speaker switching process in video conferences, abnormal switching processes are identified. Based on the video conference logs, the TCP connection establishment time is analyzed to determine if a handshake delay exists. If a handshake delay exists, the timestamps of the abnormal switching process and the handshake delay are matched to determine if the abnormal switching process is caused by the handshake delay. If so, the bandwidth data corresponding to the handshake delay during the abnormal switching process is analyzed to determine the bandwidth threshold corresponding to the handshake delay. The bandwidth change trend of the video conference switching process is monitored and analyzed to predict the timing of handshake delay adjustment. This application analyzes the time delay of the speaker switching process to mark abnormal switching processes caused by network latency. It obtains the TCP connection establishment process from the video conference logs, thereby obtaining the handshake delay time and determining if a handshake delay exists. The key timestamps of the abnormal switching process and the handshake delay are matched to determine if the abnormal switching process is caused by the handshake delay. If so, the bandwidth threshold is obtained based on the bandwidth corresponding to the handshake delay in historical switching processes. The bandwidth trend of the current switching process is monitored and analyzed to predict the timing of handshake delay adjustment, thus ensuring the stability and smoothness of the video conference. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 This is a flowchart of the steps of a monitoring, analysis and control method for a paperless meeting system according to the present invention; Figure 2 This is a logic diagram of the monitoring, analysis and control method for a paperless meeting system according to the present invention. Detailed Implementation

[0018] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0019] Please see Figures 1-2 As shown in the embodiment of the present invention, a monitoring, analysis and control method for a paperless meeting system includes the following steps: Step 1: Perform time delay analysis on the speaker switching process in the video conference to identify abnormal switching processes; Step one, the process of performing time delay analysis on the speaker switching process in the video conference, includes: A switching process is defined as the time from the start of the speaker switching trigger to the actual completion of the switching process. The conditions for determining the completion of the switching are: the audio and video streams of the new speaker are transmitted stably, and the audio and video streams of the old speaker are released. The system acquires all round-trip times during the switching process. Round-trip time refers to the time required for data to travel from the sender to the receiver, then back to the sender, and be confirmed by the sender. The unit is milliseconds. Excessive round-trip time can lead to audio and video synchronization problems and speech stuttering. As a quantitative indicator of end-to-end network latency, round-trip time provides a quantifiable and traceable basis for anomaly detection in the speaker switching process of video conferences. Calculate the average and standard deviation of the round-trip times N times before the speaker switching trigger time. If the standard deviation of the round-trip times N times before the speaker switching trigger time is less than the standard deviation threshold, the average round-trip time is used as the round-trip time baseline to reflect the round-trip time under normal conditions. Understandably, a small standard deviation indicates small fluctuations in round-trip time, in which case the mean can represent "typical latency in a stable network environment". If the standard deviation is large, it indicates that the network itself has jitter, in which case the baseline has no reference value. Calculate the mean and standard deviation of the round-trip time during the handover process, and calculate the fluctuation coefficient of the round-trip time by proportionally dividing the standard deviation by the mean. The difference between each round-trip time in the handover process and the round-trip time baseline is calculated to obtain the round-trip time change sequence. Each element in the change sequence is compared with the abnormal threshold. If it is greater than the abnormal threshold, it is marked as an abnormal element; otherwise, it is marked as a normal element. The number of intervals between anomalous elements in the round-trip time change sequence is counted, and the ratio of this number to the upper limit of the number of intervals is calculated to obtain the anomalous interval ratio. If two adjacent abnormal elements are not consecutive, it means there is a gap of one time. It should be noted that the maximum number of intervals is the total number of abnormal elements - 1; The total number of normal elements that are between abnormal elements is counted, and the ratio of this number to the number of intervals between abnormal elements is calculated to obtain the mean of abnormal intervals. The ratio of the mean of abnormal intervals to the total number of elements in the changed sequence is then calculated to obtain the ratio of the number of abnormal intervals. The abnormal continuity value of round-trip time is obtained by summing the ratio of the number of abnormal intervals and the ratio of the number of abnormal intervals. In step one, the process of identifying abnormal handover includes: The fluctuation coefficient and abnormal continuous value of the round-trip time are compared with the corresponding preset fluctuation threshold and preset continuous threshold, respectively. If the fluctuation coefficient of the round-trip time is greater than the preset fluctuation threshold and the abnormal continuous value is less than the preset continuous threshold, the handover process is marked as an abnormal handover process.

[0020] It is understandable that the physical meaning of the mean of abnormal intervals is as follows: it quantifies the continuity and density of network latency anomalies during the speaker switching process in a video conference. The lower the abnormal continuity value, the more concentrated and intermittent the abnormal elements are in the time series, that is, the network jitter exhibits the characteristics of "continuous density". This may lead to problems such as audio and video stuttering and synchronization loss during speaker switching. The more discrete and intermittent the abnormal elements are in the time series, that is, the network jitter exhibits the characteristics of "intermittent sparsity". The latency fluctuations during the switching process are more random and have a relatively mild impact on audio and video quality. The purpose of identifying abnormal handover processes is: It identifies abnormal handover processes caused by network latency, transforms the subjectively perceived "lag" into a quantitative indicator of round-trip time, filters the baseline by standard deviation and instantaneous noise by abnormal continuous values, improves detection reliability, avoids human misjudgment, and provides clear abnormal samples and time markers for subsequent steps such as handshake delay analysis and bandwidth threshold determination, forming a closed loop of "detection-analysis-optimization". Step 2: Analyze the TCP connection establishment time based on the video conference logs to determine if there is a handshake delay; The process of determining whether a handshake delay exists includes: Obtain key timestamp data of the three-way handshake during the TCP connection establishment process from the video conferencing client logs, and calculate the handshake time; Key timestamps include the timestamp T1 when the client sends the SYN packet, the timestamp T2 when the server receives the SYN packet and returns the SYN-ACK packet, the timestamp T3 when the client receives the SYN-ACK packet and sends the ACK packet, and the timestamp T4 when the server receives the ACK packet and completes the handshake. Calculate the duration of each stage: SYN transmission duration: If the client sends the SYN packet after the application layer triggers the request, the duration is T1 - application request time; Network transmission duration (SYN→SYN-ACK): T2-T1, reflecting the one-way network duration from the client to the server; Server processing time: The time difference between the server receiving SYN and sending SYN-ACK. If the time interval between T2 and the server receiving SYN is large, it may be due to high server CPU load or a full connection queue. ACK confirmation duration (SYN-ACK→ACK): T3-T2, reflecting the network time from the server to the client; Complete handshake duration: T4-T1, which is the total time from when the client initiates the handshake to when the server confirms its completion; The handshake time is compared with the standard handshake time. If the handshake time is greater than or equal to the standard handshake time, then there is a handshake delay. Conversely, if the handshake delay is less than the standard handshake time value, it indicates that the handshake delay is within the normal range and will not affect the smoothness of the meeting. The standard value for handshake time is set by those skilled in the art; For example, in the log of an abnormal handover process, it was found that the client sent an STN packet at T0+0.2s, received the SYN-ACK packet from the server at T0+0.6s, and completed the ACK confirmation at T0+0.8s. This handshake time reached 600ms (0.8s-0.2s), which exceeded the 300ms threshold, and it was determined that there was a handshake delay. TCP is a connection-oriented, reliable transport layer protocol. To ensure accurate and orderly data transmission, a three-way handshake is required between the client and server to establish a connection before data is actually sent. The specific process is as follows: First handshake (SYN): The client sends a TCP segment with the SYN (synchronization sequence number) flag to the server. This segment contains an initial sequence number randomly generated by the client (let's say x). This operation indicates that the client is requesting to establish a connection with the server and informing itself of the starting sequence number to be used for subsequent data transmission. The second handshake (SYN-ACK): After receiving the SYN packet from the client, the server will return a packet with SYN and ACK (acknowledgment) flags. The SYN flag is used to synchronize the server's initial sequence number (let's say it's y), and the ACK flag is used to acknowledge the client's request. The acknowledgment number is x+1, indicating that the server has received the client's SYN packet and is ready to establish a connection. The third handshake (ACK): After receiving the SYN-ACK message from the server, the client sends a message with the ACK flag to the server, with the acknowledgment number y+1, indicating that the client has received the synchronization information from the server. At this point, the connection between the two parties is officially established and data transmission can begin. Through a three-way handshake, the client and server confirm each other's existence and synchronize their sequence numbers, laying the foundation for reliable data transmission in the future. The handshake time refers to the time taken from when the client sends a SYN packet to when the server finally receives the client's ACK packet, which is the entire three-way handshake process. In an ideal network environment, the TCP handshake latency is relatively low. It is understandable that the purpose of determining whether there is a handshake delay during the handover process is: In video conferencing speaker switching scenarios, excessively long handshake times may prevent the new speaker's media stream from being established in a timely manner, resulting in audio and video stuttering, black screens, and other issues that affect the smoothness of the meeting and user experience. Analyzing and optimizing TCP handshake latency can help ensure the stable operation of video conferencing. Step 3: If a handshake delay exists, timestamp the abnormal handover process and the handshake delay to determine whether the abnormal handshake process was caused by the handshake delay. The process of determining whether an abnormal handover is caused by a handshake delay includes: Obtain key timestamps of the abnormal switching process, including: switching trigger time: the moment when the application layer initiates the switching request; and actual switching completion time: the moment when the new stream is stable and the old stream is released. Obtain key timestamps for handshake delay, including: T1: time when the client sends the SYN packet, T2: time when the server returns the SYN-ACK packet, T3: time when the client sends the ACK packet, and T4: time when the server receives the ACK and completes the handshake. The abnormal switching time window is [T tr ,T com ], where T tr To switch trigger times, T com To switch the actual completion time, the handshake delay time window is [T1, T4]; Calculate the overlap duration between the abnormal handover window and the handshake delay window. Overlap duration = max(0, min(T4, T)). com )-max(T1,T tr )), where T tr To switch trigger times, T com To switch between actual completion times, calculate window overlap. The window overlap is compared with the overlap threshold. If the window overlap is greater than or equal to the overlap threshold, it indicates that the abnormal handover and handshake delay are highly correlated in time, and the next step is performed. Verify the timing logic by matching the handshake delay window with the abnormal handover window using key timestamps. If the first handshake delay phase (T2-T1) is within t0ms after the handover is initiated, and the handshake completion time T4 is before the handover completion time, then determine whether the abnormal handover process is caused by the handshake delay. The setting of t0 is as follows: the human senses are usually sensitive to audio and video delays at around t0. If the delay exceeds t0ms, users may perceive stuttering in speech and desynchronization between video and audio, which will affect the meeting experience. Step 4: If so, analyze the bandwidth data corresponding to the handshake delay during the abnormal handover process to determine the bandwidth threshold corresponding to the handshake delay; In step four, the process of determining the bandwidth threshold corresponding to the handshake delay includes: Obtain the bandwidth corresponding to the handshake delay during the historical handover process, generate a bandwidth sequence, calculate the mean and standard deviation of the bandwidth sequence, and calculate the bandwidth fluctuation amplitude by proportionally dividing the standard deviation of the bandwidth sequence by the mean. The bandwidth fluctuation amplitude is compared with the bandwidth threshold. If it is less than or equal to the bandwidth threshold, the mean of the bandwidth sequence is used as the bandwidth threshold; otherwise, the minimum value in the bandwidth sequence is used as the bandwidth threshold. It is understandable that the purpose of setting a bandwidth threshold is: By calculating the mean and standard deviation of the bandwidth sequence corresponding to the handshake delay during historical handover, the typical level and fluctuation range of network bandwidth are quantified. If the bandwidth fluctuation range is small, it indicates that the network is stable, and the mean is directly used as the bandwidth threshold to represent the "baseline bandwidth under normal working conditions". If the fluctuation range is large, it indicates that the network is prone to extreme situations, and the minimum value of the sequence is used as the bandwidth threshold to ensure that a safety buffer can still be reserved when the bandwidth drops sharply. Step 5: Monitor and analyze the bandwidth change trend during the video conference switching process, and predict the timing of handshake delay by combining the bandwidth threshold corresponding to the handshake delay.

[0021] The prediction process for the handshake delay includes: The bandwidth of the current video conference is monitored in real time, a bandwidth threshold is set, and the current bandwidth is compared with the bandwidth threshold. If the current bandwidth exceeds the bandwidth threshold, the difference between the current bandwidth and the bandwidth threshold is calculated to obtain the approximate value. The bandwidth threshold is set by those skilled in the art. The purpose of setting the bandwidth threshold is to provide the system with a time window for analysis and prediction by identifying abnormal traffic trends in advance, and ultimately to mitigate the risk before handshake delay occurs through control measures. The approach value is compared with the approach threshold. If the approach value is less than or equal to the approach threshold, the time point when the current bandwidth reaches the bandwidth threshold is marked as the critical time point. The time period between the critical time point and the current time point is marked as the over-limit period. The bandwidth threshold values ​​within the over-limit period are integrated according to the time sequence to obtain the critical bandwidth sequence. The correlation coefficient of the critical bandwidth sequence is calculated based on the Pearson correlation coefficient method. If the correlation coefficient is greater than or equal to the correlation threshold, the critical bandwidth sequence is linear; if the correlation coefficient is less than the correlation threshold, the critical bandwidth sequence is nonlinear. If the critical bandwidth sequence is linear, the least squares method is used for linear fitting. The fitting process is as follows: Let the trend equation be F(t) = kt + b, where t is time, k is the slope, and b is the intercept. Substituting the critical bandwidth sequence into the equation and solving for k and b, we can obtain the predicted time when the bandwidth reaches the bandwidth threshold. , where F thre This is the bandwidth threshold; If the critical bandwidth sequence is nonlinear, random forest is used to predict the time when the bandwidth reaches the bandwidth threshold. The prediction process is as follows: Construct a feature matrix, i.e., input variables, including: real-time bandwidth, critical bandwidth sequence [F(t-1), F(t-2), ..., F(tn)], where n is the number of bandwidths, and the rate of change feature: first-order difference of bandwidth. Bandwidth second-order difference ; The construction logic of decision trees is as follows: Split Criteria: Using mean squared error (MSE) as the node splitting criterion, we select features and thresholds that minimize the variance of the child node sample labels. Random subspace: Each decision tree is randomly selected from the feature matrix during splitting. Evaluate each feature (d is the total number of features) to reduce inter-tree correlation; Sample sampling: Bootstrap sampling (sampling with replacement) was used, with approximately 63.2% of the training samples used per tree, and the remaining samples used for out-of-bag (OOB) error evaluation; Key parameter settings: Number of trees: Determine the optimal value through grid search (usually 100-200 trees); Maximum depth: Limits the depth of the tree to avoid overfitting, usually set to 5-10 layers; Minimum number of sample splits: The minimum number of samples required for node splitting, set to 10-20 (to avoid the influence of noise). The specific prediction process is as follows: The input variable feature values ​​are normalized to the [0,1] interval. Each tree performs forward inference on the input vector to obtain the prediction time of a single tree. The final prediction time is the average of the prediction values ​​of all trees, which is the time when the handshake delay occurs.

[0022] The advantages of the technical solution in this application are as follows: This application analyzes the time delay of speaker switching in video conferences to identify abnormal switching processes. Based on video conference logs, it analyzes the TCP connection establishment time to determine if handshake delay exists. If handshake delay is present, it matches the timestamps of the abnormal switching process and the handshake delay to determine if the abnormal switching is caused by the handshake delay. If so, it analyzes the bandwidth data corresponding to the handshake delay during the abnormal switching process to determine the corresponding bandwidth threshold. It also monitors and analyzes the bandwidth change trend during the video conference switching process to predict the timing for handshake delay adjustment. By analyzing the time delay of speaker switching, this application identifies abnormal switching processes caused by network latency. It obtains the TCP connection establishment process from the video conference logs to obtain the handshake delay time, determines if handshake delay exists, matches the key timestamps of the abnormal switching process and the handshake delay to determine if the abnormal switching is caused by the handshake delay, and if so, obtains the bandwidth threshold based on the bandwidth corresponding to the handshake delay in historical switching processes. Finally, it monitors and analyzes the bandwidth trend of the current switching process to predict the timing for handshake delay adjustment, thereby ensuring the stability and smoothness of video conferences.

[0023] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A monitoring, analysis, and control method for a paperless conferencing system, characterized in that: include: Step 1: Perform time delay analysis on the speaker switching process in the video conference to identify abnormal switching processes; The method for identifying abnormal switching processes is as follows: Obtain all round-trip times during the handover process, calculate the mean and standard deviation of the round-trip times, and calculate the fluctuation coefficient of the round-trip times by proportionally dividing the standard deviation by the mean; calculate the abnormal continuous values ​​of the round-trip times, and compare the fluctuation coefficient of the round-trip times with the preset fluctuation threshold and the abnormal continuous values ​​with the preset continuous threshold. If the fluctuation coefficient of the round-trip time is greater than the preset fluctuation threshold and the abnormal continuous value is less than the preset continuous threshold, the handover process will be marked as an abnormal handover process. Step 2: Analyze the TCP connection establishment time based on the video conference logs to determine if there is a handshake delay; Step 3: If a handshake delay exists, timestamp the abnormal handover process and the handshake delay to determine whether the abnormal handshake process was caused by the handshake delay. Step 4: If so, analyze the bandwidth data corresponding to the handshake delay during the abnormal handover process to determine the bandwidth threshold corresponding to the handshake delay; Step 5: Monitor and analyze the bandwidth change trend during the video conference switching process, and predict the timing of handshake delay by combining the bandwidth threshold corresponding to the handshake delay.

2. The monitoring, analysis, and control method for a paperless conference system according to claim 1, characterized in that: The calculation steps for the abnormal continuous values ​​include: The difference between each round-trip time in the switching process and the round-trip time baseline is calculated, and the calculation results are integrated according to the time sequence to obtain the change sequence of round-trip time. For each element in the change sequence, if the element does not meet the requirements, it is marked as an abnormal element; otherwise, it is marked as a normal element. The number of intervals between anomalous elements in the round-trip time change sequence is counted, and the ratio of this number to the upper limit of the number of intervals is calculated to obtain the anomalous interval ratio. The maximum number of intervals is the total number of abnormal elements minus 1. The total number of normal elements that are between abnormal elements is counted, and the ratio of this number to the number of intervals between abnormal elements is calculated to obtain the mean of abnormal intervals. The ratio of the mean of abnormal intervals to the total number of elements in the changed sequence is then calculated to obtain the ratio of the number of abnormal intervals. The abnormal continuity value of round-trip time is obtained by summing the ratio of the number of abnormal intervals and the ratio of the number of abnormal intervals.

3. The monitoring, analysis, and control method for a paperless conference system according to claim 1, characterized in that: The process for determining whether a handshake delay exists is as follows: Obtain the key timestamp data of the three-way handshake in the TCP connection establishment process, and get the time of the three handshakes. The handshake duration is the sum of the three handshake times. The handshake time is compared with the standard handshake time value. If the handshake time is greater than or equal to the standard time value, then there is a handshake delay.

4. The monitoring, analysis, and control method for a paperless conference system according to claim 1, characterized in that: The method for determining whether an abnormal handover process is caused by a handshake delay is as follows: Obtain key timestamps of the abnormal handover process, including the handover trigger time, the actual handover completion time, and key timestamps of handshake delay, including the sending and receiving timestamps of the three-way handshake; The abnormal handover time window and handshake delay time window are calculated based on the key timestamp; Calculate the overlap duration of the abnormal handover time window and the handshake delay time window, and calculate the window overlap degree by proportionally calculating the abnormal handover window duration. If the window overlap is greater than or equal to the overlap threshold, and the handshake completion time is before the handover completion time, then the abnormal handover process is caused by handshake delay.

5. The monitoring, analysis, and control method for a paperless conference system according to claim 4, characterized in that: The abnormal handover time window is from the handover trigger time to the actual handover completion time, and the handshake delay time window is from the first handshake sending time to the third handshake receiving time.

6. The monitoring, analysis, and control method for a paperless conference system according to claim 1, characterized in that: The method for obtaining the bandwidth threshold corresponding to the handshake delay is as follows: Obtain the bandwidth corresponding to the handshake delay during the historical handover process, generate a bandwidth sequence, calculate the mean and standard deviation of the bandwidth sequence, and calculate the bandwidth fluctuation amplitude by proportionally dividing the standard deviation of the bandwidth sequence by the mean. The bandwidth fluctuation amplitude is compared with the bandwidth threshold. If it is less than or equal to the bandwidth threshold, the mean of the bandwidth sequence is used as the bandwidth threshold; otherwise, the minimum value in the bandwidth sequence is used as the bandwidth threshold.

7. The monitoring, analysis, and control method for a paperless conference system according to claim 6, characterized in that: The method for predicting the timing of the handshake delay is as follows: The critical bandwidth sequence is obtained by comparing the current bandwidth with the set bandwidth threshold, and the correlation coefficient of the critical bandwidth sequence is calculated based on the Pearson correlation coefficient method. The correlation coefficient of the critical bandwidth sequence is compared with the preset correlation coefficient. If the correlation coefficient meets the requirements, the critical bandwidth sequence is linear; otherwise, the critical bandwidth sequence is nonlinear. The timing of handshake delay can be predicted using linear and nonlinear critical bandwidth sequences.

8. The monitoring, analysis, and control method for a paperless conference system according to claim 7, characterized in that: The critical bandwidth sequence is obtained as follows: The bandwidth of the current video conference is monitored in real time, a bandwidth threshold is set, and the current bandwidth is compared with the bandwidth threshold. If the current bandwidth exceeds the bandwidth threshold, the difference between the current bandwidth and the bandwidth threshold is calculated to obtain the approximate value. The approach value is compared with the approach threshold. If the approach value is less than or equal to the approach threshold, the time point when the current bandwidth reaches the bandwidth threshold is marked as the critical time point. The time period between the critical time point and the current time point is marked as the over-limit period. The bandwidth threshold values ​​within the over-limit period are integrated in time sequence to obtain the critical bandwidth sequence.

9. The monitoring, analysis, and control method for a paperless conference system according to claim 7, characterized in that: The method for predicting the handshake delay time using a linear critical bandwidth sequence is as follows: If the critical bandwidth is linear, the least squares method is used to linearly fit the critical bandwidth sequence to obtain the handshake delay prediction model. The bandwidth threshold is used as input, and the handshake delay prediction model outputs the predicted time point when the bandwidth reaches the bandwidth threshold, which is the time when the handshake delay occurs.

10. The monitoring, analysis, and control method for a paperless conference system according to claim 7, characterized in that: The method for predicting the handshake delay time using a nonlinear critical bandwidth sequence is as follows: Random forests are used to predict the time when bandwidth reaches the bandwidth threshold, and a feature matrix is ​​constructed as input variables, including real-time bandwidth and critical bandwidth sequences. The input variables are normalized to the [0,1] interval. Each decision tree in the random forest performs forward inference on the input variables to obtain the prediction time of a single decision tree. The final prediction time is the average of the predictions of all decision trees, which is the time when the handshake delay occurs.

Images