A cloud phone automated operation and maintenance method based on WeChat robot

Abstract

The present invention discloses a cloud phone automated operation and maintenance method based on WeChat robot, which relates to the field of cloud phone automated operation and maintenance technology, and includes the following steps: for all devices, an initial high segmentation density is set, and the mobile phone image file is transmitted in the set initial small-capacity blocks; during the initial transmission of the mobile phone image file, the block density is gradually adjusted, and the segmentation density is reduced in a gradient increase manner, so that the amount of data in each block gradually increases. The present invention uses a machine learning model to perceive and analyze the processing power of the device in real time, intelligently classify and dynamically optimize the transmission strategy, ensure the normal operation of the device, and avoid system failures and data loss. By dynamically adjusting the block density, hierarchical transmission is performed for devices with different processing capabilities, and device resources are efficiently utilized, transmission efficiency is improved, system stability is enhanced, and the risks brought by device differences are effectively addressed.

Description

Technical Field

[0001] The present invention relates to the technical field of cloud phone automated operation and maintenance, and in particular to a cloud phone automated operation and maintenance method based on a WeChat robot. Background Art

[0002] Automated cloud phone operation and maintenance based on WeChat robots refers to the use of WeChat robot technology to realize the automation of cloud phone operation and maintenance management through the WeChat platform. The WeChat robot developed through the Wechaty framework receives operation and maintenance instructions, and then interacts with the background interface to automatically perform various cloud phone operation and maintenance tasks, such as remote execution of Shell or Python scripts to manage devices, batch operations, update built-in applications, monitor device status, and share custom images. In this way, operation and maintenance personnel can send commands directly on WeChat, and the system will automatically handle complex operation and maintenance tasks and feedback the results in real time, greatly improving the efficiency and response speed of operation and maintenance.

[0003] During the automated operation and maintenance of cloud phones, custom image sharing is instructed through WeChat robots. Operation and maintenance personnel can manage and distribute cloud phone image files to achieve rapid deployment and environmental recovery of multiple devices (the devices here are cloud phone devices, i.e. virtualized mobile devices) or regions. The custom image sharing process is as follows: WeChat robots capture and parse the shared image instructions, extract key parameters (such as image ID, target device, region, etc.), and then communicate with the cloud phone backend through HTTP API to send a push request. The backend server finds the corresponding file based on the image ID and pushes it to the target device or region. After the push is completed, the system feedbacks the results (such as success status, number of devices, etc.) to the WeChat group through the WeChat robot for the operation and maintenance personnel to view in real time.

[0004] In the prior art, when distributing cloud phone image files to multiple devices, in order to improve the distribution efficiency, the image files are usually simply split before transmission. However, due to differences in the processing capabilities of devices, the risk of network transmission interruption or delay increases significantly when devices with weaker processing capabilities receive larger split blocks. These devices often cannot keep up with the transmission speed of larger data blocks, resulting in frequent failures in the transmission process. Even if some data is successfully transmitted, devices with insufficient processing power may not be able to process the received data in a timely manner, resulting in cache overflow, which in turn causes data loss or errors. Ultimately, these problems will result in the inability to fully assemble the image file, causing the device to fail to operate normally or the image function to be limited, seriously affecting the overall stability and reliability of the system.

[0005] 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

[0006] The purpose of the present invention is to provide a cloud phone automated operation and maintenance method based on WeChat robot, which intelligently perceives the real-time data processing status of the device through a machine learning model, accurately predicts the device processing capacity, automatically performs classification and optimization processing, responds to device status changes in real time, prevents system failures or data transmission failures caused by insufficient processing capacity, ensures the normal operation of the device and data integrity, dynamically adjusts the block density, performs hierarchical processing for devices with different processing capabilities, efficiently optimizes the data transmission strategy, avoids cache overflow and data loss, improves data transmission efficiency and system stability, makes full use of high-processing capacity devices to accelerate transmission, and ensures smooth transmission of low-processing capacity devices, so as to solve the problems in the above-mentioned background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a cloud phone automated operation and maintenance method based on WeChat robot, comprising the following steps:

[0008] For all devices, set the initial high segmentation density and transfer the mobile phone image file in the set initial small capacity blocks;

[0009] During the initial transmission of the mobile phone image file, the block density is gradually adjusted, and the segmentation density is reduced by a gradient increase method, so that the amount of data in each block gradually increases;

[0010] During the data transmission process, the data receiving information of each device when receiving the image file is obtained in real time, and the actual performance of the device during the data transmission process is perceived;

[0011] Extract key features from the data received by each device, perform abnormal analysis on the extracted features within a set fixed-time monitoring window, use a pre-trained machine learning model to conduct a comprehensive analysis on the features after abnormal analysis, and perceive the real-time situation of each device when processing image files;

[0012] Based on the predictions of the machine learning model, the processing power of each device is classified and evaluated. According to the prediction results of the machine learning model, the devices are divided into two categories: high processing power devices and low processing power devices;

[0013] For devices with high processing power, continue to use the gradient method to increase the amount of data in each block, further reduce the block density, and increase the amount of data in each block;

[0014] For devices with low processing power, an optimized transmission strategy is adopted. Specifically, the segmentation density is increased based on the device's ability to process mobile phone image files, the amount of data in each block is reduced, and the segmentation density of mobile phone image files is dynamically optimized according to the device's real-time data processing situation to ensure that the device can efficiently transmit mobile phone image files.

[0015] Preferably, reducing the segmentation density by a gradient increase means that during the initial transmission of the mobile phone image file, the size of each data block is gradually adjusted so that the volume of the data block gradually increases, while the number of blocks decreases accordingly.

[0016] Preferably, key features are extracted from the data receiving information of each device, wherein the extracted features include the frequency of errors occurring when the device reassembles the image file after receiving the data and the frequency of the device buffer being filled with data. After acquisition, within a set fixed-duration monitoring window, the frequency of errors occurring when the device reassembles the image file after receiving the data and the frequency of the device buffer being filled with data are subjected to abnormal analysis and processing, and a data reconstruction error quantization reference value and a cache overflow quantization reference value are generated respectively. Through the data reconstruction error quantization reference value and the cache overflow quantization reference value, the potential processing capacity abnormality of each device when processing the image file is intelligently perceived.

[0017] Preferably, after obtaining the data reconstruction error quantization reference value and the cache overflow quantization reference value generated by each device when processing the image file, the data reconstruction error quantization reference value and the cache overflow quantization reference value are comprehensively analyzed using a pre-trained machine learning model to generate a processing capacity coefficient, and the processing capacity coefficient is used to intelligently predict the processing capacity of the current device when processing the image file.

[0018] Preferably, the processing capacity coefficient generated when each device receives the image file is compared and analyzed with a preset processing capacity coefficient reference threshold by a machine learning model under a fixed-duration monitoring window, and the processing capacity of each device is divided. The specific division process is as follows:

[0019] If the processing capacity coefficient is greater than or equal to the processing capacity coefficient reference threshold, the device is classified as a high processing capacity device under the fixed-duration monitoring window;

[0020] If the processing capability coefficient is less than the processing capability coefficient reference threshold, the device is classified as a low processing capability device in the fixed-duration monitoring window.

[0021] Preferably, within a set fixed-duration monitoring window, the specific steps of generating a data reorganization error quantitative reference value after performing abnormal analysis on the frequency of errors that occur when the device reorganizes the image file after receiving data are as follows:

[0022] Divide the fixed-duration monitoring window into

[0023] nIn the whole process from receiving data to reassembling the image file, the number of reassembly errors of the device in each time window is monitored in real time. The weight of the reassembly error in each time window is calculated based on the number of reassembly errors. The calculation expression is:

[0024] , where is the number of reorganization errors within the time window, is the total number of packets processed by the device in this time window, is the processing delay of the device at the current moment, It is a nonlinear adjustment factor used to dynamically adjust the weight according to the delay to ensure that the differences in processing capabilities of different devices are taken into account. is the recombination error weight calculated within each time window;

[0025] A dynamic change coefficient is introduced to reflect the change of processing performance. The dynamic change coefficient considers the response speed of the device in different time windows and is defined as:

[0026] , where is the dynamic change coefficient, is the response time of the device within the time window, is the acceleration factor of the response time, reflecting the impact of the device processing speed on the reorganization error. is the length of the time window, Indicates the duration of the fixed-duration monitoring window;

[0027] The dynamic change coefficient of the device and the reorganization error weight are combined to finally generate the data reorganization error quantization reference value of the device. The expression for generating the data reorganization error quantization reference value is:

[0028] , where Quantify reference values ​​for data reconstruction errors, To adjust the nonlinear coefficient of exponential sensitivity, It is an adjustment factor that controls the contribution of the recombination error weight in the reference value of data recombination error quantification.

[0029] Preferably, within a set fixed-duration monitoring window, the specific steps of generating a buffer overflow quantitative reference value after abnormal analysis of the frequency of the device buffer being filled with data are as follows:

[0030] In the set fixed-time monitoring window, the device's cache overflow is preliminarily counted. When the cache overflows, the overflow time is recorded. , and accumulate the number of times each overflow occurs. The counting expression is: , where represents the cache overflow frequency function at time t, for Dirac Function, when the buffer overflows ,otherwise , N Indicates the total number of overflows within the set monitoring window; calculates the cumulative rate of device cache overflows. Considering the load caused by data accumulation and cache failure to be cleaned up in time, the time derivative of the cache overflow frequency is introduced to obtain the cumulative rate of overflows. The calculation expression of the cumulative rate is:

[0031] , where represents the cumulative rate of buffer overflows, is the decay factor of the overflow time, the decay rate Controls the weight variation of spillovers over time, Indicates the integral variable, representing an intermediate point in time, used to calculate a certain moment from the starting time 0 to the current time t;

[0032] By further analyzing the cumulative rate change of cache overflow, combining device performance and cache response speed, a cache overflow quantitative reference value is generated. The specific expression is:

[0033]

[0034] In the formula, It is the final generated buffer overflow quantitative reference value, reflecting the severity of the buffer overflow problem handled by the device. T is the total duration of the fixed-duration monitoring window.

[0035] Preferably, for devices with low processing power, an optimized transmission strategy is used to increase the segmentation density. The specific steps of dynamically optimizing the segmentation density of the mobile phone image file are as follows:

[0036] In the current fixed-duration monitoring window, first obtain the processing capacity coefficient of the low-processing capacity device in real time, and calculate the segmentation density adjustment coefficient based on the current processing capacity coefficient of the device and the processing capacity coefficient reference threshold. The calculation expression of the segmentation density adjustment coefficient is:

[0037] , where is the processing capacity coefficient of the device under the current fixed-duration monitoring window, is the segmentation density adjustment factor, is the reference threshold of the processing capacity coefficient, It is a tuning parameter that controls the intensity of adjusting the split density of the mobile phone image file on low-processing devices. The adjustment amplitude has a nonlinear characteristic, and the adjustment is more obvious when the processing capacity coefficient is less than the processing capacity coefficient reference threshold or the processing capacity coefficient is higher than the processing capacity coefficient reference threshold;

[0038] Based on the segmentation density adjustment coefficient, the actual segmentation density under the current fixed-duration monitoring window is dynamically adjusted. When the segmentation density of the mobile phone image file increases, the amount of data in each block decreases. The calculation expression for dynamic adjustment is: , is the segmentation density under the current fixed-time monitoring window, is the actual segmentation density in the next fixed-length monitoring window. hour, , indicating the need to increase the actual segmentation density and reduce the amount of data in each block. hour, , which means reducing the actual segmentation density and increasing the amount of data in each block.

[0039] Preferably, for low processing power devices, when the processing power coefficient is again greater than or equal to the processing power coefficient reference threshold, the gradient increase method will no longer be used to further reduce the block density and thereby increase the amount of data in each block. Instead, the current segmentation density will be dynamically adjusted.

[0040] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0041] The present invention uses a machine learning model to intelligently perceive and analyze the real-time data processing of the device, accurately predict the processing capacity of each device, and automatically classify and optimize the devices. This intelligent classification and dynamic optimization strategy can respond to the state changes of the device in real time, effectively prevent system failures or data transmission failures caused by insufficient device processing capacity, and ensure the normal operation of the device and the integrity of the image file. At the same time, combined with real-time data analysis and anomaly detection, the system can adjust the transmission strategy for low-processing power devices in a targeted manner, thereby ensuring the efficient operation and stability of the entire cloud phone operation and maintenance system, and reducing the risks caused by device differences.

[0042] The present invention greatly improves the efficiency of data transmission by dynamically adjusting the data transmission strategy of block density and performing hierarchical processing for devices with different processing capabilities. For high-processing capability devices, a gradient increase method is used to reduce the segmentation density, increase the size of data blocks, and fully utilize the high processing capability of the device to accelerate transmission; while for low-processing capability devices, an optimized transmission strategy is used to dynamically adjust the block density and reduce the size of data blocks, so that data can be efficiently transmitted under limited processing capabilities. This not only avoids problems such as cache overflow and data loss caused by insufficient processing capability of the device during the transmission process, but also ensures the overall stability of the system and significantly improves data transmission efficiency and system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] 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.

[0044] Figure 1 The present invention is a method flow chart of a cloud phone automated operation and maintenance method based on WeChat robot. DETAILED DESCRIPTION

[0045] 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.

[0046] The present invention provides Figure 1 The method for automated operation and maintenance of a cloud phone based on a WeChat robot includes the following steps:

[0047] For all devices, set the initial high segmentation density and transfer the mobile phone image file in the set initial small capacity blocks;

[0048] High split density initial chunking means that the image file will be split into smaller chunks so that each device can easily receive and process the initial chunks. High split density helps reduce the risk of transmission failure, especially on devices with weaker processing power. Smaller data chunks can avoid data backlogs and cache overflows caused by slow processing speeds.

[0049] During the initial transmission of the mobile phone image file, the block density is gradually adjusted, and the segmentation density is reduced by a gradient increase method, so that the amount of data in each block gradually increases;

[0050] The gradient increase method is used to reduce the segmentation density, which means that during the initial transmission of the mobile phone image file, the size of each data block is gradually adjusted, so that the volume of the data block gradually increases, while the number of blocks decreases accordingly.

[0051] In the initial stage of transmission, in order to ensure that all devices (regardless of their processing power) can receive data smoothly, the system will split the mobile phone image file into multiple smaller data blocks. This means that the segmentation density is high, that is, the number of data blocks is large and the amount of data in each data block is small. This can reduce the processing pressure of a single device and avoid cache overflow or data loss caused by too large data blocks.

[0052] As the transmission proceeds smoothly, the system will gradually reduce the segmentation density through a gradient increase, which is manifested in increasing the amount of data in each data block. This means merging or enlarging the original small data blocks to increase the unit data volume of each transmission. For example, each data block may only be a few MB at first, and after several rounds of transmission, the size of the data block may increase to tens of MB or even larger.

[0053] The gradient increase method means that during the data block size adjustment process, the system will not increase the data block size significantly at one time, but will increase it gradually and in a gradient manner. For example, the data block size can be gradually increased from the initial 5MB to 10MB, 20MB, 50MB, etc. This adjustment follows a gradual increase process, rather than jumping directly from 5MB to 50MB. This gradual adjustment ensures that the device has enough time to process slightly increased data volume at each stage, thereby avoiding the processing pressure caused by the sudden increase in data blocks.

[0054] Larger data blocks can reduce the overhead between blocks and the number of transmission control packets, thereby improving overall transmission efficiency. For devices with strong processing capabilities, larger data blocks can transmit large amounts of data more quickly. Although smaller data blocks are easier to process, too many divisions will increase the total transmission time. When data blocks are gradually enlarged, the amount of data transmitted each time increases, which can reduce the time waiting for the next data block to arrive, thereby reducing overall transmission delays. For devices with high processing capabilities, gradually enlarging data blocks can better utilize their hardware resources, enabling them to process image files in a more efficient manner and complete transmission quickly.

[0055] During the data transmission process, the data receiving information of each device when receiving the image file is obtained in real time, and the actual performance of the device during the data transmission process is perceived;

[0056] During the data transmission process, obtaining the data reception information of each device when receiving the image file is a key step. The performance of the device can be monitored in real time in a variety of ways. First, the system can monitor the transmission rate, packet loss rate and retransmission of each device through the transmission control protocol (TCP) at the network layer. The TCP protocol can reflect the network status and the basic transmission performance of the device when receiving data. Second, the device can feedback its internal resource information such as CPU usage, memory occupancy and cache status. This data can help identify whether the device has performance bottlenecks when processing data. Third, through the network layer log and system layer log, you can track the details of the device when receiving and processing data packets, such as delay, packet loss, error code, etc. Fourth, using traffic analysis tools (such as NetFlow or sFlow), you can monitor the data traffic pattern and transmission path of the device to evaluate the network load and the real-time performance of the device. In addition, the application layer protocol on the device side can also provide feedback. The application can reflect the reception quality and processing of the image file by reporting the number of received file blocks, processing time, decoding errors and other information. In these ways, the system can fully perceive the performance of the device during data transmission and ensure timely adjustment of the transmission strategy.

[0057] Extract key features from the data received by each device, perform abnormal analysis on the extracted features within a set fixed-time monitoring window, use a pre-trained machine learning model to conduct a comprehensive analysis on the features after abnormal analysis, and perceive the real-time situation of each device when processing image files;

[0058] Key features are extracted from the data receiving information of each device, wherein the extracted features include the frequency of errors when the device reassembles the image file after receiving the data and the frequency of the device buffer being filled with data. After acquisition, within a set fixed-duration monitoring window, the frequency of errors when the device reassembles the image file after receiving the data and the frequency of the device buffer being filled with data are analyzed and processed abnormally, and a data reconstruction error quantization reference value and a cache overflow quantization reference value are generated respectively. The data reconstruction error quantization reference value and the cache overflow quantization reference value are used to intelligently perceive the potential processing capacity abnormality of each device when processing the image file.

[0059] After obtaining the data reconstruction error quantization reference value and the cache overflow quantization reference value generated by each device when processing the image file, the data reconstruction error quantization reference value and the cache overflow quantization reference value are comprehensively analyzed using a pre-trained machine learning model to generate a processing capacity coefficient. The processing capacity coefficient is used to intelligently predict the processing capacity of the current device when processing the image file.

[0060] A pre-trained machine learning model is one that is trained using a large amount of historical data before the system goes live, so that it can learn and identify performance patterns and anomalies when the device processes data. This model establishes a complex relationship between input and output by learning input features (such as the frequency of reorganization errors, cache overflow frequency, transmission rate, etc. when the device processes data) and output results (such as device processing capabilities, failure rates, etc.) from historical data. The model can improve the accuracy of predictions by continuously optimizing algorithm parameters such as weights and biases. During the training process, the model is repeatedly adjusted until its predictions of device performance reach the ideal level. This training process usually uses supervised learning methods, which means that the model is trained using a data set with known labels, where the labels can represent the device's processing capabilities or anomalies.

[0061] In actual applications, the trained machine learning model is applied to real-time data analysis. When the data reorganization error quantization reference value and buffer overflow quantization reference value generated by the device are input into the model, the model will perform a comprehensive analysis of these new data based on the patterns it learned during the training process. Through this comprehensive analysis, the model can generate a processing capacity coefficient to evaluate the current performance of the device when processing image files. Because the model has learned the performance of the device under different loads and states, it can make intelligent predictions about the processing capacity of the current device. This prediction can not only identify whether the current device is in normal working condition, but also provide early warning of potential performance bottlenecks or failures. By leveraging pre-trained machine learning models, the system can achieve more accurate device performance evaluation and optimization during data transmission and processing.

[0062] A high frequency of errors when the device reassembles the image file after receiving data usually indicates that the device has poor processing capabilities. Errors when reassembling the image file mean that the device encountered problems when processing the received data blocks. This may be due to insufficient CPU, memory, or storage resources on the device, which prevents it from processing the data in a timely and correct manner, resulting in frequent data reassembly errors. This situation usually reflects that the device's computing power, data processing efficiency, or system resource allocation is insufficient and cannot withstand the current data processing load. Therefore, a high frequency of reassembly errors is one of the important indicators of poor device processing capabilities.

[0063] The specific steps for generating a quantitative reference value for data reassembly errors after abnormal analysis and processing of the frequency of errors that occur when the device reassembles the image file after receiving data within the set fixed-time monitoring window are as follows:

[0064] Divide the fixed-duration monitoring window into n time windows of equal length (where nIndicates the total number of time windows in a fixed-length monitoring window). During the entire process from device data reception to image file reassembly, the number of reassembly errors that occur on the device in each time window is monitored in real time. The reassembly error weight in each time window is calculated based on the number of reassembly errors. The calculation expression is:

[0065] , where is the number of reorganization errors within the time window, is the total number of packets processed by the device in this time window, is the processing delay of the device at the current moment (the sum of transmission delay and processing delay), It is a nonlinear adjustment factor used to dynamically adjust the weight according to the delay to ensure that the differences in processing capabilities of different devices are taken into account. is the recombination error weight calculated within each time window;

[0066] Specifically, Dynamically changes according to the specific latency of the device (such as transmission latency or processing latency). When the latency of some devices is large, will increase, thus giving more weight adjustment to devices with larger latency, so that the difference in processing capabilities of devices with higher latency can be better reflected in the formula. The value is kept small to reduce its impact on the calculation. The values ​​in different scenarios are not specifically limited here and can be set according to actual conditions and needs.

[0067] The processing capacity of a device may fluctuate over time, so a dynamic change coefficient is introduced to reflect the change in processing performance. The dynamic change coefficient considers the response speed of the device in different time windows and is defined as:

[0068] , where is the dynamic change coefficient, is the response time of the device within the time window, is the acceleration factor of the response time, reflecting the impact of the device processing speed on the reorganization error. is the length of the time window, Indicates the duration of the fixed-duration monitoring window;

[0069] Within this time window, the response time of the device refers to the total time it takes for the device to receive the data packet from the beginning to complete the processing and feedback the result. Specifically, the response time includes the transmission delay during the data reception process, the calculation time for the device to process the data internally, and the time it takes for the device to feedback the processing result. It is a key indicator that reflects the overall efficiency of the device when receiving and processing image files. The shorter the response time, the higher the performance of the device when processing data; on the contrary, a longer response time usually indicates that the device may have a resource bottleneck or a heavy processing load, resulting in a decrease in processing capacity. This step obtains the dynamic changes in processing capacity by comprehensively calculating the reorganization errors and response times of the device in different time windows.

[0070] The dynamic change coefficient of the device and the reorganization error weight are combined to finally generate the data reorganization error quantization reference value of the device. The expression for generating the data reorganization error quantization reference value is:

[0071] , where Quantify reference values ​​for data reconstruction errors, To adjust the nonlinear coefficient of exponential sensitivity, It is an adjustment factor that controls the contribution of the recombination error weight in the reference value of data recombination error quantification.

[0072] It is used to adjust the sensitivity of the data reorganization error quantization reference value to the dynamic change coefficient of the equipment processing capacity. When calculating the processing capacity of the device, you need to consider the fluctuation of the device's processing capacity and the fault tolerance of the entire system. If the processing capacity of the device fluctuates greatly in different time windows, It should be set to a higher value to reduce the excessive impact of dynamic changes in equipment on the final index, so that the system can have a higher tolerance for short-term anomalies, thereby reducing unnecessary alarms and misjudgments. On the contrary, if a rapid response to changes in equipment processing capacity is required, especially in mission-critical scenarios, It can be set to a lower value to make the system more sensitive to changes in device performance and adjust the transmission strategy in a timely manner. The value of is determined through experiments or historical data analysis based on the expected fluctuation range of the device processing capability and the fault tolerance requirements in the application scenario.

[0073] It is used to control the contribution of the data reassembly error weight in the data reassembly error quantification reference value, so it determines the impact of the severity of the error in the data reassembly process on the overall device processing capability evaluation. When you use the image file, you need to consider the sensitivity of the device to reconstruction errors when processing the image file. If the device is very sensitive to reconstruction errors (for example, data corruption will seriously affect subsequent operations), It should be set to a higher value to increase the weight of the reassembly error on the total index, so that the system pays more attention to the data integrity during processing. On the contrary, if the device can tolerate a certain degree of reassembly errors and these errors can be compensated by mechanisms such as retransmission, It can be set to a lower value to reduce its impact on the overall index. Set it through experiments or according to specific application requirements , which can ensure that the factor is more reasonable and effective in evaluating the system performance.

[0074] It can be seen from the data reassembly error quantization reference value that, within the set fixed-time monitoring window, the data reassembly error quantization reference value generated after abnormal analysis of the frequency of errors when the device reassembles the image file after receiving data is processed, the larger the performance value, the higher the frequency of errors when the device reassembles the image file within the set fixed-time monitoring window, which means that the device encounters more problems during data processing, reflecting its poor data processing capability. A high performance value means that the device may not be able to effectively process the received data blocks due to insufficient resources, slow processing speed or frequent errors, resulting in a decrease in overall performance. On the contrary, if the data reassembly error quantization reference value is small, it means that the device can process data stably and correctly, and there are fewer errors during the reassembly process, indicating that the device has a strong data processing capability. Therefore, the size of the data reassembly error quantization reference value directly reflects the current data processing capability of the device.

[0075] When a device receives a mobile phone image file, a high frequency of the device's cache being filled with data usually indicates that the device has poor processing capabilities. A full cache means that the device cannot process the received data in a timely manner, resulting in new data being unable to enter the cache, causing data accumulation. This may be due to insufficient resources such as the device's CPU and memory, which cannot keep up with the speed of data processing, or the data processing algorithm is inefficient. When the cache overflows frequently, the device is likely to experience data loss, delays, or transmission interruptions, further affecting the overall performance of the device and the transmission and reorganization of image files. Therefore, a high frequency of cache filling is an obvious signal that there is a bottleneck in the device's ability to process data under the current load.

[0076] The specific steps for generating a buffer overflow quantitative reference value after abnormal analysis of the frequency of the device buffer being filled with data within the set fixed-time monitoring window are as follows:

[0077] In the set fixed-time monitoring window, the device's cache overflow is preliminarily counted. When the cache overflows, the overflow time is recorded. , and accumulate the number of times each overflow occurs. The counting expression is: , where represents the cache overflow frequency function at time t, for Dirac Function, when the buffer overflows ,otherwise , N Indicates the total number of overflows within the set monitoring window;

[0078] Through this step, the time point and frequency of each cache overflow during the monitoring period can be obtained, and the state of cache overflow when data accumulates can be captured.

[0079] The cumulative rate of device cache overflow is calculated. Considering the load caused by data accumulation and cache failure to be cleaned up in time, the time derivative of cache overflow frequency is introduced to obtain the cumulative rate of overflow. The calculation expression of the cumulative rate is:

[0080] , where represents the cumulative rate of buffer overflows, is the decay factor of the overflow time, the decay rate The weight of the control overflow changes with time. The overflow in the distant time has less impact on the current one. In this formula, represents an integral variable, which represents an intermediate point in time and is used to calculate a certain moment from the starting time 0 to the current time t. Specifically, It is an intermediate time variable used in the integration process to calculate the frequency of buffer overflow during this time period. The contribution of . The integral expression Represents the cumulative overflow frequency from the past time 0 to the current time t. As time t advances, the past moments The influence of control);

[0081] This step captures the dynamic changes of the cache overflow accumulation rate through time integration and combined with the decay factor, focusing on measuring the impact of the overflow on subsequent transmissions.

[0082] By further analyzing the cumulative rate change of cache overflow, combining device performance and cache response speed, a cache overflow quantitative reference value is generated. The specific expression is:

[0083] , where It is the final generated buffer overflow quantitative reference value, reflecting the severity of the buffer overflow problem handled by the device. T is the total duration of the fixed-duration monitoring window (same as above Indicates that the fixed-duration monitoring windows have the same duration).

[0084] It can be seen from the buffer overflow quantitative reference value that the larger the performance value of the buffer overflow quantitative reference value generated after abnormal analysis of the frequency of the device buffer being filled with data within the set fixed-time monitoring window, the higher the frequency of the device buffer being filled with data within the set fixed-time monitoring window, which means that the device cannot process the received data in time and the cache frequently overflows, which reflects that the data processing capability of the device is poor. Conversely, if the buffer overflow quantitative reference value is small, it means that the device can efficiently process the transmitted data and the buffer overflow is less, indicating that the data processing capability of the device is good. Therefore, the buffer overflow quantitative reference value can be used as an important indicator of the performance of the device when processing data, and the processing capability of the device can be judged by monitoring its changes.

[0085] The machine learning model is not specifically limited here, and can realize the data reorganization error quantification reference value and buffer overflow quantization reference value Perform comprehensive analysis to generate processing capacity coefficients In order to realize the technical solution of the present invention, the present invention provides a specific implementation method; the processing capacity coefficient The resulting calculation formula is:

[0086] , where , Quantify the reference values ​​for data reconstruction errors and buffer overflow quantization reference value The preset scaling factor of , Both are greater than 0.

[0087] The preset proportionality coefficient here indicates the weight of the influence of the data reorganization error quantization reference value and the buffer overflow quantization reference value on the final result when calculating the processing capacity coefficient. The size of the preset proportionality coefficient reflects how the system balances the contribution of the two indicators when evaluating the device processing capacity. For example, The influence weight of the control data reorganization error quantification reference value, while Control the influence weight of the buffer overflow quantization reference value. By adjusting these preset proportional coefficients, the calculation results of the processing power can be optimized according to the needs of the actual application scenario, making the model more flexible and accurate in reflecting the performance of the device.

[0088] It can be seen from the processing capacity coefficient that, within the set fixed-duration monitoring window, the larger the performance value of the data reconstruction error quantization reference value generated after the abnormal analysis and processing of the frequency of errors when the device reassembles the image file after receiving the data, the larger the performance value of the cache overflow quantization reference value generated after the abnormal analysis and processing of the frequency of the device buffer area being filled with data, that is, the smaller the performance value of the processing capacity coefficient generated by the machine learning model when each device receives the image file under the fixed-duration monitoring window, the worse the data processing capability of the device, and vice versa.

[0089] Based on the predictions of the machine learning model, the processing power of each device is classified and evaluated. According to the prediction results of the machine learning model, the devices are divided into two categories: high processing power devices and low processing power devices;

[0090] The processing capacity coefficient generated by each device when receiving the image file is compared and analyzed with the preset processing capacity coefficient reference threshold by the machine learning model in a fixed-duration monitoring window, and the processing capacity of each device is divided. The specific division process is as follows:

[0091] If the processing capacity coefficient is greater than or equal to the processing capacity coefficient reference threshold, the device is classified as a high processing capacity device under the fixed-duration monitoring window. A high processing capacity device refers to a device that can process larger blocks and has a higher transmission rate, and is suitable for continuing to transmit large amounts of data.

[0092] If the processing capacity coefficient is less than the processing capacity coefficient reference threshold, the device is classified as a low processing capacity device under the fixed-duration monitoring window. Low processing capacity devices may encounter problems when receiving larger blocks due to slow processing speed and unstable network conditions.

[0093] The classified devices will adopt different transmission strategies to ensure that each device can complete the image file reception efficiently and securely.

[0094] For devices with high processing power, continue to use the gradient method to increase the amount of data in each block, further reduce the block density, and increase the amount of data in each block;

[0095] The strategy of gradually increasing the amount of data in each block can maximize the transmission efficiency, allowing these devices to complete the reception of image files at a faster speed. By gradually increasing the amount of data and reducing the number of blocks, high-processing power devices can quickly complete the image file transmission without affecting processing performance.

[0096] For devices with low processing power, an optimized transmission strategy is adopted. Specifically, the segmentation density is increased based on the device's ability to process mobile phone image files, the amount of data in each block is reduced, and the segmentation density of mobile phone image files is dynamically optimized according to the device's real-time data processing situation to ensure that the device can efficiently transmit mobile phone image files.

[0097] For devices with low processing power, an optimized transmission strategy is used to increase the segmentation density. The specific steps for dynamically optimizing the segmentation density of mobile phone image files are as follows:

[0098] In the current fixed-duration monitoring window, first obtain the processing capacity coefficient of the low-processing capacity device in real time, and calculate the segmentation density adjustment coefficient based on the current processing capacity coefficient of the device and the processing capacity coefficient reference threshold. The calculation expression of the segmentation density adjustment coefficient is:

[0099] , where is the processing capacity coefficient of the device under the current fixed-duration monitoring window, is the segmentation density adjustment factor, is the reference threshold of the processing capacity coefficient, It is an adjustment parameter that controls the intensity of adjusting the split density of mobile phone image files on devices with low processing power. It is not specifically limited here and can be set comprehensively according to actual conditions and needs. The adjustment amplitude has a nonlinear characteristic, and the adjustment is more obvious when the processing capacity coefficient is less than the processing capacity coefficient reference threshold or the processing capacity coefficient is higher than the processing capacity coefficient reference threshold;

[0100] Based on the segmentation density adjustment coefficient, the actual segmentation density under the current fixed-duration monitoring window is dynamically adjusted. When the segmentation density of the mobile phone image file increases, the amount of data in each block decreases. The calculation expression for dynamic adjustment is: , is the segmentation density under the current fixed-time monitoring window, is the actual segmentation density in the next fixed-length monitoring window. hour, , indicating the need to increase the actual segmentation density and reduce the amount of data in each block. hour, , which means reducing the actual segmentation density and increasing the amount of data in each block;

[0101] It should be noted that for low processing power devices, when the processing power coefficient is again greater than or equal to the processing power coefficient reference threshold, the gradient increase method will no longer be considered to further reduce the block density and increase the amount of data in each block. Instead, the current segmentation density will be dynamically adjusted.

[0102] This step can increase the segmentation density (i.e., reduce the data block size) when the device's processing capability is poor, ensuring that the device can better process data; when the device's processing capability is restored or better, the system will increase the segmentation density to improve transmission efficiency.

[0103] The present invention uses a machine learning model to intelligently perceive and analyze the real-time data processing of the device, accurately predict the processing capacity of each device, and automatically classify and optimize the devices. This intelligent classification and dynamic optimization strategy can respond to the state changes of the device in real time, effectively prevent system failures or data transmission failures caused by insufficient device processing capacity, and ensure the normal operation of the device and the integrity of the image file. At the same time, combined with real-time data analysis and anomaly detection, the system can adjust the transmission strategy for low-processing power devices in a targeted manner, thereby ensuring the efficient operation and stability of the entire cloud phone operation and maintenance system, and reducing the risks caused by device differences.

[0104] The present invention greatly improves the efficiency of data transmission by dynamically adjusting the data transmission strategy of block density and performing hierarchical processing for devices with different processing capabilities. For high-processing capability devices, a gradient increase method is used to reduce the segmentation density, increase the size of data blocks, and fully utilize the high processing capability of the device to accelerate transmission; while for low-processing capability devices, an optimized transmission strategy is used to dynamically adjust the block density and reduce the size of data blocks, so that data can be efficiently transmitted under limited processing capabilities. This not only avoids problems such as cache overflow and data loss caused by insufficient processing capability of the device during the transmission process, but also ensures the overall stability of the system and significantly improves data transmission efficiency and system performance.

[0105] 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.

[0106] 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.

[0107] 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 cloud phone automated operation and maintenance method based on WeChat robot, characterized in that: The following steps are involved: For all devices, set the initial high segmentation density and transfer the mobile phone image file in the set initial small capacity blocks; During the initial transmission of the mobile phone image file, the block density is gradually adjusted, and the segmentation density is reduced by a gradient increase method, so that the amount of data in each block gradually increases; The gradient increase method is used to reduce the segmentation density. It means that during the initial transmission of the mobile phone image file, the size of each data block is gradually adjusted, so that the size of the data block gradually increases, while the number of blocks decreases accordingly. During the data transmission process, the data receiving information of each device when receiving the image file is obtained in real time, and the actual performance of the device during the data transmission process is perceived; Extract key features from the data received by each device, perform abnormal analysis on the extracted features within a set fixed-time monitoring window, use a pre-trained machine learning model to conduct a comprehensive analysis on the features after abnormal analysis, and perceive the real-time situation of each device when processing image files; Extract key features of the data receiving information of each device, wherein the extracted features include the frequency of errors when the device reassembles the image file after receiving the data and the frequency of the device buffer being filled with data. After acquisition, within a set fixed-time monitoring window, perform abnormal analysis on the frequency of errors when the device reassembles the image file after receiving the data and the frequency of the device buffer being filled with data, and generate data reassembly error quantization reference values ​​and buffer overflow quantization reference values ​​respectively. Through the data reassembly error quantization reference values ​​and buffer overflow quantization reference values, intelligent perception of potential processing capacity abnormalities of each device when processing image files is performed; Based on the predictions of the machine learning model, the processing power of each device is classified and evaluated. According to the prediction results of the machine learning model, the devices are divided into two categories: high processing power devices and low processing power devices; For devices with high processing power, continue to use the gradient method to increase the amount of data in each block, further reduce the block density, and increase the amount of data in each block; For devices with low processing power, an optimized transmission strategy is adopted. Specifically, the segmentation density is increased based on the device's ability to process mobile phone image files, the amount of data in each block is reduced, and the segmentation density of mobile phone image files is dynamically optimized according to the device's real-time data processing situation to ensure that the device can efficiently transmit mobile phone image files.

2. According to claim 1, a cloud phone automated operation and maintenance method based on WeChat robot is characterized in that: After obtaining the data reconstruction error quantization reference value and the cache overflow quantization reference value generated by each device when processing the image file, the data reconstruction error quantization reference value and the cache overflow quantization reference value are comprehensively analyzed using a pre-trained machine learning model to generate a processing capacity coefficient. The processing capacity coefficient is used to intelligently predict the processing capacity of the current device when processing the image file.

3. According to claim 2, a cloud phone automated operation and maintenance method based on WeChat robot is characterized in that: The processing capacity coefficient generated by each device when receiving the image file is compared and analyzed with the preset processing capacity coefficient reference threshold by the machine learning model in a fixed-duration monitoring window, and the processing capacity of each device is divided. The specific division process is as follows: If the processing capacity coefficient is greater than or equal to the processing capacity coefficient reference threshold, the device is classified as a high processing capacity device under the fixed-duration monitoring window; If the processing capability coefficient is less than the processing capability coefficient reference threshold, the device is classified as a low processing capability device in the fixed-duration monitoring window.

4. According to claim 1, a cloud phone automated operation and maintenance method based on WeChat robot is characterized in that: The specific steps for generating a quantitative reference value for data reassembly errors after abnormal analysis and processing of the frequency of errors that occur when the device reassembles the image file after receiving data within the set fixed-time monitoring window are as follows: Divide the fixed-duration monitoring window into n In the whole process from receiving data to reassembling the image file, the number of reassembly errors of the device in each time window is monitored in real time. The weight of the reassembly error in each time window is calculated based on the number of reassembly errors. The calculation expression is: , where is the number of reorganization errors within the time window, is the total number of packets processed by the device in this time window, is the processing delay of the device at the current moment, It is a nonlinear adjustment factor used to dynamically adjust the weight according to the delay to ensure that the differences in processing capabilities of different devices are taken into account. is the recombination error weight calculated within each time window; A dynamic change coefficient is introduced to reflect the change of processing performance. The dynamic change coefficient considers the response speed of the device in different time windows and is defined as: , where is the dynamic change coefficient, is the response time of the device within the time window, is the acceleration factor of the response time, reflecting the impact of the device processing speed on the reorganization error. is the length of the time window, Indicates the duration of the fixed-duration monitoring window; The dynamic change coefficient of the device and the reorganization error weight are combined to finally generate the data reorganization error quantization reference value of the device. The expression for generating the data reorganization error quantization reference value is: , where Quantify reference values ​​for data reconstruction errors, To adjust the nonlinear coefficient of exponential sensitivity, It is an adjustment factor that controls the contribution of the recombination error weight in the reference value of data recombination error quantification.

5. According to claim 1, a cloud phone automated operation and maintenance method based on WeChat robot is characterized in that: The specific steps for generating a buffer overflow quantitative reference value after abnormal analysis of the frequency of the device buffer being filled with data within the set fixed-time monitoring window are as follows: In the set fixed-time monitoring window, the device's cache overflow is preliminarily counted. When the cache overflows, the overflow time is recorded. , and accumulate the number of times each overflow occurs. The counting expression is: , where represents the cache overflow frequency function at time t, for Dirac Function, when the buffer overflows ,otherwise , N Indicates the total number of overflows within the set monitoring window; The cumulative rate of device cache overflow is calculated. Considering the load caused by data accumulation and cache failure to be cleaned up in time, the time derivative of cache overflow frequency is introduced to obtain the cumulative rate of overflow. The calculation expression of the cumulative rate is: , where represents the cumulative rate of buffer overflows, is the decay factor of the overflow time, the decay rate Controls the weight variation of spillovers over time, Indicates the integral variable, representing an intermediate point in time, used to calculate a certain moment from the starting time 0 to the current time t; By further analyzing the cumulative rate change of cache overflow, combining device performance and cache response speed, a cache overflow quantitative reference value is generated. The specific expression is: , where It is the final generated buffer overflow quantitative reference value, reflecting the severity of the buffer overflow problem handled by the device. T is the total duration of the fixed-duration monitoring window.

6. According to claim 3, a cloud phone automated operation and maintenance method based on WeChat robot is characterized in that: For devices with low processing power, an optimized transmission strategy is used to increase the segmentation density. The specific steps for dynamically optimizing the segmentation density of mobile phone image files are as follows: In the current fixed-duration monitoring window, first obtain the processing capacity coefficient of the low-processing capacity device in real time, and calculate the segmentation density adjustment coefficient based on the current processing capacity coefficient of the device and the processing capacity coefficient reference threshold. The calculation expression of the segmentation density adjustment coefficient is: , where is the processing capacity coefficient of the device under the current fixed-duration monitoring window, is the segmentation density adjustment factor, is the reference threshold of the processing capacity coefficient, It is a tuning parameter that controls the intensity of adjusting the split density of the mobile phone image file on low-processing devices. The adjustment amplitude has a nonlinear characteristic, and the adjustment is more obvious when the processing capacity coefficient is less than the processing capacity coefficient reference threshold or the processing capacity coefficient is higher than the processing capacity coefficient reference threshold; Based on the segmentation density adjustment coefficient, the actual segmentation density under the current fixed-duration monitoring window is dynamically adjusted. When the segmentation density of the mobile phone image file increases, the amount of data in each block decreases. The calculation expression for dynamic adjustment is: , is the segmentation density under the current fixed-time monitoring window, is the actual segmentation density in the next fixed-length monitoring window. hour, , indicating the need to increase the actual segmentation density and reduce the amount of data in each block. hour, , which means reducing the actual segmentation density and increasing the amount of data in each block.

7. According to claim 1, a cloud phone automated operation and maintenance method based on WeChat robot is characterized in that: For low processing power devices, when the processing power coefficient is again greater than or equal to the processing power coefficient reference threshold, the gradient increase method will no longer be used to further reduce the block density and increase the amount of data in each block. Instead, the current segmentation density will be dynamically adjusted.

Images