An integrated management system for an Internet-based data service cloud platform

Abstract

The present invention relates to the technical field of data service cloud platforms, and specifically provides a comprehensive management system for an Internet-based data service cloud platform, including: a data cleaning and classification module, a load balancing module, a cache server, a server group, a mirror backup module, a disaster recovery module, and a visualization operation and maintenance module. By cleaning and classifying data, the huge data stream is diverted, and the smooth weighted round-robin algorithm is used to perform secondary diversion on the huge data stream. The sub-data stream is preferentially allocated to the servers with good load conditions, and the data of the server group of other data service cloud platforms is backed up. Sensors are used to judge whether a disaster signal is sent to improve the backup efficiency. At the same time, the overall load condition of the data service cloud platform is visualized, and the real-time data service ability is optimized based on the underlying architecture of the data center, providing an interface service integrating visualization, multi-services, and panoramic operation.

Description

Technical Field

[0001] The present invention relates to the technical field of data service cloud platforms, and particularly to an integrated management system for an Internet-based data service cloud platform. Background Art

[0002] A data service cloud platform is an Internet-based basic service for business data, which provides a bridge between a data warehouse, a computing engine, and standardized data. The cloud platform helps enterprises achieve data sharing, data integration, and data services, and solves problems such as data being unable to be shared, difficult to use, and difficult to access. The data service cloud platform integrates various data storage applications and analysis services of large enterprises into a unified environment, and can provide enterprises with an available, trustworthy, and dynamically updated data service mechanism, as well as service supports such as data management and business process integration. The functions of the data service cloud platform include high-efficiency distributed processing, real-time processing, real-time streaming accounting, etc. However, while the data service cloud platform provides comprehensive data services for enterprises, there are also some deficiencies. For example, in some cases, such as storage-intensive services like log retention and traffic lists, the CPU usage rate is long-term lower than the ideal level. At the same time, although computing services consume a high amount of CPU, the utilization rate of storage resources is very low. This indicates that there may be irrationalities in the resource allocation and use of the platform, resulting in a large amount of resources being idle; at the same time, with the increase in data volume and the diversification of data types, the data service platform faces greater challenges in data security and privacy protection. The integrated management of the data service cloud platform needs to ensure the efficient, safe, and stable operation of the cloud platform while meeting the data service needs of users. Therefore, it is necessary to design an integrated management system for an Internet-based data service cloud platform to solve the above technical problems. Summary of the Invention

[0003] The purpose of the present invention is to provide an integrated management system for an Internet-based data service cloud platform to solve the problems mentioned in the above background art.

[0004] The purpose of the present invention can be achieved through the following technical solutions:

[0005] An integrated management system for an Internet-based data service cloud platform includes: a data cleaning and classification module, a load balancing module, a cache server, a server group, a mirror backup module, a disaster recovery module, and a visualization operation and maintenance module.

[0006] Furthermore, the data cleaning and classification module is used to clean and classify the data stream input into the cloud platform. Before the cleaning starts, the original data is backed up, and then the data stream is split into more than two sub-data streams. The documents and annotations in each sub-data stream are read to determine the information type and data type of the sub-data stream.

[0007] The information types include text information, image information, video and audio information, user data, structured data, real-time data streams, geographical location data, application data, etc. Each information type corresponds to a preset information type factor a.

[0008] The data types include integer type, floating-point type, list, array, hash table, etc. Each data type corresponds to a preset data type factor b.

[0009] Through the formula the caching factor of each sub-data stream is obtained. When is greater than or equal to the preset caching factor threshold, the sub-data stream is output to the cache server; when is less than the preset caching factor threshold, the sub-data stream is output to the load balancing module.

[0010] Furthermore, the load balancing module distributes the sub-data streams to each server in the server group through the smooth weighted round-robin algorithm, combines the analysis of the load conditions of the servers to distribute the sub-data streams, avoids the downtime caused by excessive load on a single server, and at the same time simulates the adverse impact on the load of a single server after the influx of sub-data streams through the algorithm to dynamically adjust the weights. It not only realizes the priority distribution of data streams to the servers with better load conditions, but also prevents a large amount of data streams from continuously flowing into the servers with better load conditions, ensuring that the entire server group realizes high-efficiency cooperation through reasonable task distribution as a whole; at the same time, it also reduces the number of self-load detections of each server and reduces unnecessary computing power waste. The specific process is as follows:

[0011] In the first step, the load balancing module sends a heartbeat packet signal to all servers at preset intervals. After receiving the heartbeat packet signal, server i reads the CPU occupancy rate and the disk read rate , where i is the server number, i = 1, 2,..., n; n = the total number of servers. Through the formula the load weight is obtained, and is sent to the load balancing module;

[0012] In the second step, the load balancing module counts all , and sends the first sub-data stream Data1 to the server with the largest . When the largest comes from multiple servers, Data1 is sent to the server with the smallest i among them. The of the server that receives Data1 is marked as ;

[0013] In the third step, all are adjusted. The adjustment process is as follows: First, process the marked , let ; Remove the mark, and restore to ; Then, process other unmarked , let ;

[0014] Step 4, the load balancing module counts all , and sends the second sub-data stream Data2 to the largest server. When the largest comes from multiple servers, send Data2 to the server with the smallest i among them, and mark the of the server that receives Data2 as , and return to Step 3.

[0015] For example, when there are three servers A, B, and C in the server group, the corresponding load weights are , , , judge the largest is , send Data1 to server A, and then adjust = 5 - (5 + 1 + 1) + 5 = 3, = +1 + 1 = 2, = 1 + 1 = 2, judge the largest is , send Data2 to server A; similarly, after adjustment = 1, = 3, = 3, and so on. Finally, the allocated servers for data streams 1, 2, 3, 4, and 5 are A, A, B, A, and C.

[0016] Further, the cache server is used to store data streams with more access times, uses a disk with faster read and write speeds to meet a large number of data retrieval operations sent by the server. After receiving the data stream, the cache server records the CPU occupancy rate , disk read rate and memory occupancy rate every preset time, where p is the time number, p = 1, 2,..., q; q is the time number period. When p = q, let the next p = 1 to start the next recording period. Send p, , and to the visualization and operation and maintenance module.

[0017] Further, each server in the server group jointly undertakes the data processing task through the task allocation of the load balancing module. After receiving the sub-data stream, the cache server records the CPU occupancy rate every preset time , disk read rate and memory occupancy rate , where p is the time number, p = 1, 2,..., q; q is the time number period. When p = q, let the next p = 1 to start the next recording cycle. Send p, , and to the visualization and operation and maintenance module

[0018] Further, the visualization and operation and maintenance module records p, , , , , and , and performs data visualization processing to analyze the operation of the server group and the cache server, and feeds back the analysis results to the operation and maintenance personnel. Obtain the operation parameters of the server through the formula , where , , and are preset weight factors; obtain the operation parameters of the cache server through the formula , where , , and are preset weight factors

[0019] Take the time number p as the abscissa, as the ordinate to establish a two-dimensional rectangular coordinate system. Fill all points (p, ) into the coordinate system to obtain the two-dimensional discrete random variable distribution diagram of the operation parameters of the server changing with the time number. Connect all points with a smooth curve, and then fit the smooth curve with an elementary function image to obtain the change function image changing with p. Use the fast Fourier transform to transform the image from the time domain to the frequency domain to obtain the server operation parameter spectrum diagram. Compare it with the standard server operation parameter spectrum diagram. The peak points in the server operation parameter spectrum diagram are recorded as ( , ), and the valley points are recorded as ( , ), where u is the number of peak points, u = 1, 2, 3, ..., v; v is the total number of peak points; o is the number of valley points, o = 1, 2, 3, ..., s; s is the total number of valley points. The peak points in the standard server operation parameter spectrum are denoted as ( , ), and the valley points are denoted as ( , ). Here, u is the number of the peak point, u = 1, 2, 3, ..., v; v is the total number of peak points; o is the number of the valley point, o = 1, 2, 3, ..., s; s is the total number of valley points. Through the formula the anomaly degree parameter of the server group is obtained.

[0020] Taking the time number p as the abscissa, as the ordinate to establish a two-dimensional rectangular coordinate system, and filling all points (p, ) into the coordinate system to obtain the two-dimensional discrete random variable distribution diagram of the operation parameters of the cache server changing with the time number. Connect all points with a smooth curve, and then fit the smooth curve with an elementary function image to obtain the change function image changing with p. Using the fast Fourier transform to transform the image from the time domain to the frequency domain, the operation parameter spectrum diagram of the cache server is obtained. Comparing it with the standard operation parameter spectrum diagram of the cache server, the peak points in the operation parameter spectrum diagram of the cache server are denoted as ( , ), and the valley points are denoted as ( , ). Here, u is the number of peak points, u = 1, 2, 3, ..., v; v is the total number of peak points; o is the number of valley points, o = 1, 2, 3, ..., s; s is the total number of valley points. The peak points in the standard operation parameter spectrum diagram of the cache server are denoted as ( , ), and the valley points are denoted as ( , ). Here, u is the number of the peak point, u = 1, 2, 3, ..., v; v is the total number of peak points; o is the number of the valley point, o = 1, 2, 3, ..., s; s is the total number of valley points. Through the formula the anomaly degree parameter of the cache server is obtained.

[0021] Through the formula the comprehensive anomaly parameter is obtained. The comprehensive anomaly parameter Reflects the operation status of the server group and cache servers. When is higher than the preset comprehensive anomaly parameter threshold, it indicates that the overall operation load of the server group is too large. Several reference handling methods will be provided to the operation and maintenance personnel. Method 1: Increase the number of servers in the server group; Method 2: Optimize the algorithm of the load balancing module; Method 3: Update the hardware devices in the servers.

[0022] When the server group receives a disaster signal from this data service cloud platform, it immediately stops the operations in all servers and quickly compresses and sends the data to the mirror backup module.

[0023] Furthermore, the mirror backup module is responsible for periodically storing some data in other data service cloud platforms during daily operations. Multiple mirror backup modules, as a whole, completely preserve all the data of all other platforms. A single mirror backup module periodically mirrors and copies some data in the server groups of other data service cloud platforms, formats the copied data every preset time, and then accesses other data service cloud platforms again to mirror and copy the data in each server. When the mirror backup module receives a disaster signal, it immediately copies the data of the server group of the data service cloud platform where the disaster signal comes from, and incorporates the servers with operating parameters less than the operating parameter threshold into the mirror backup module.

[0024] Furthermore, the disaster recovery module is responsible for sending disaster signals to the mirror backup module during power outages, fires, floods, and other natural disasters. The disaster recovery module includes a power outage sensor, a smoke sensor, and a seismic sensor. When the data obtained by the power outage sensor, the smoke sensor, or the seismic sensor exceeds the threshold, it sends a disaster signal to the server group and all the mirror backup modules of other data service cloud platforms, and issues a warning message to the operation and maintenance management personnel through the display screen.

[0025] Advantages of the present invention:

[0026] (1) Shunt the huge data stream according to the data type, perform targeted processing on different types of data, and use the smooth weighted round-robin algorithm to perform secondary shunting on the data and send it to different servers, avoiding the problems of too high CPU occupancy rate and resource idleness, optimizing the resource utilization efficiency of the cloud platform, and ensuring the high efficiency and stability of the comprehensive management of the data service cloud platform;

[0027] (2) Back up the information in the servers through the mirror backup module and the disaster recovery module, make timely responses in case of adverse disasters, perform efficient backup of the data in the server group, ensure the recoverability of the data, and improve the data security of the comprehensive management of the data service cloud platform;

[0028] (3) Visualize and fuse the load conditions of the server to facilitate the optimization of real-time data service capabilities, ensuring that the data service cloud platform integrated management system constructs an interface business capability integrating visualization, multi-services, and panoramic operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The present invention will be further described below with reference to the accompanying drawings.

[0030] Figure 1 is a schematic diagram of the connection of system modules of the present invention;

[0031] Figure 2 is a schematic diagram of the working method of the mirror backup module of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0033] As Figure 1 shown, a data service cloud platform integrated management system based on the Internet includes: a data cleaning and classification module, a load balancing module, a cache server, a server group, a mirror backup module, a disaster recovery module, and a visualization and operation and maintenance module.

[0034] The data cleaning and classification module is used to clean and classify the data stream input into the cloud platform. Before the cleaning starts, the original data is backed up, and then the data stream is split into more than two sub-data streams. The documents and annotations in each sub-data stream are read to determine the information type and data type of the sub-data stream.

[0035] The information types include text information, image information, video and audio information, user data, structured data, real-time data streams, geographical location data, application program data, etc. Each information type corresponds to a preset information type factor a.

[0036] The data types include integer type, floating point type, list, array, hash table, etc. Each data type corresponds to a preset data type factor b.

[0037] Through the formula the cache factor of each sub-data stream is obtained. When is greater than or equal to the preset cache factor threshold, the sub-data stream is output to the cache server; when When the value is less than the preset cache factor threshold, the sub-data stream is output to the load balancing module.

[0038] The load balancing module distributes sub-data streams to each server in the server group through a smooth weighted polling algorithm, distributes sub-data streams based on the load of the server, and avoids downtime caused by excessive load on a single server. At the same time, the algorithm simulates the adverse effects of the influx of sub-data streams on the load of a single server, and dynamically adjusts the weights. It not only realizes the priority allocation of data streams to servers with better load conditions, but also prevents huge amounts of data streams from continuously pouring into servers with better load conditions, ensuring that the entire server group as a whole can achieve efficient collaboration through reasonable task allocation; it also reduces the number of self-load detections of each server and reduces unnecessary waste of computing power. The specific process is:

[0039] In the first step, the load balancing module sends a heartbeat packet signal to all servers at a preset time interval. After receiving the heartbeat packet signal, server i reads the CPU usage and disk read rate , where i is the server number, i=1,2,...,n; n=total number of servers. The values ​​of CPU usage and disk read rate are calculated by the formula Get load weight ,Will Sent to the load balancing module; λ1 and λ2 are preset weight factors;

[0040] The second step is to load balance all , send the first sub-data stream Data1 to The largest server, when the largest When Data1 comes from multiple servers, send Data1 to the server with the smallest i, and Mark as ;

[0041] Step 3: Adjust all , the adjustment process is: first, process the marked ,make ; Unmark, Restore to ; Then, process the other unmarked ,make ;

[0042] Step 4: The load balancing module counts all , send the second sub-data stream Data2 to The largest server, when the largest When it comes from multiple servers, send Data2 to the server with the smallest i, and mark the server that receives Data2 as and return to step 3.

[0043] For example, when there are three servers A, B, and C in the server group, and the corresponding load weights are , , , judge that the largest is , send Data1 to server A, and then adjust =5 - (5 + 1 + 1) + 5 = 3, = +1 + 1 = 2, =1 + 1 = 2, judge that the largest is , send Data2 to server A; similarly, after adjustment =1, =3, =3, and so on. Finally, the allocated servers for data streams 1, 2, 3, 4, and 5 are A, A, B, A, and C.

[0044] The cache server is used to store data streams with a large number of access times, uses a disk with a faster read and write speed to meet a large number of data retrieval operations sent by the server. After receiving the data stream, the cache server records the CPU occupancy rate , disk read rate and memory occupancy rate every preset time, where p is the time number, p = 1, 2,..., q; q is the time number period. When p = q, let the next p = 1 and start the next recording cycle. Send p, , and to the visualization and operation and maintenance module.

[0045] Each server in the server group jointly undertakes the data processing task through the task allocation of the load balancing module. After receiving the sub-data stream, the cache server records the CPU occupancy rate , disk read rate and memory occupancy rate every preset time, where p is the time number, p = 1, 2,..., q; q is the time number period. When p = q, let the next p = 1 and start the next recording cycle. Send p, , and to the visualization and operation and maintenance module.

[0046] The visualization and operation and maintenance module records p, ​, , , , and , and perform data visualization processing, analyze the operation of the server group and cache servers, and feedback the analysis results to the operation and maintenance personnel. Through the formula obtain the operation parameters of the server , where , and are preset weight factors; through the formula obtain the operation parameters of the cache server , where , and are preset weight factors.

[0047] Taking the time number p as the abscissa and as the ordinate, establish a two-dimensional rectangular coordinate system. Fill all the points (p, ) into the coordinate system to obtain the two-dimensional discrete random variable distribution diagram of the server operation parameters changing with the time number. Connect all the points with a smooth curve, and then fit the smooth curve with an elementary function image to obtain the change function image changing with p. Use the fast Fourier transform to transform the image from the time domain to the frequency domain to obtain the server operation parameter spectrum diagram. Compare it with the standard server operation parameter spectrum diagram. The peak points in the server operation parameter spectrum diagram are denoted as ( , ), the trough points are denoted as ( , ), u is the number of peak points, u = 1, 2, 3,..., v; v is the total number of peak points; o is the number of trough points, o = 1, 2, 3,..., s; s is the total number of trough points. The peak points in the standard server operation parameter spectrum diagram are denoted as ( , ), the trough points are denoted as ( , ), u is the number of peak points, u = 1, 2, 3,..., v; v is the total number of peak points; o is the number of trough points, o = 1, 2, 3,..., s; s is the total number of trough points. Through the formula obtain the abnormality parameter of the server group.

[0048] Taking the time number p as the abscissa and as the ordinate, establish a two-dimensional rectangular coordinate system. Fill all the points (p, ) Fill in the coordinate system to obtain a two-dimensional discrete random variable distribution diagram of the operating parameters of the cache server changing with the time number. Connect all points with a smooth curve, and then fit the smooth curve with an elementary function image to obtain The change function image changing with p. Use the fast Fourier transform to transform the image from the time domain to the frequency domain to obtain the operating parameter spectrum diagram of the cache server. Compare it with the standard operating parameter spectrum diagram of the cache server. The peak points in the operating parameter spectrum diagram of the cache server are recorded as ( , ), and the trough points are recorded as ( , ), where u is the number of peak points, u = 1, 2, 3,..., v; v is the total number of peak points; o is the number of trough points, o = 1, 2, 3,..., s; s is the total number of trough points. The peak points in the standard operating parameter spectrum diagram of the cache server are recorded as ( , ), and the trough points are recorded as ( , ), where u is the number of peak points, u = 1, 2, 3,..., v; v is the total number of peak points; o is the number of trough points, o = 1, 2, 3,..., s; s is the total number of trough points. Through the formula Obtain the anomaly degree parameter of the cache server .

[0049] Through the formula Obtain the comprehensive anomaly parameter , where K1 and k2 are preset weight factors. The comprehensive anomaly parameter Reflects the operating conditions of the server group and the cache server. When is higher than the preset comprehensive anomaly parameter threshold, it indicates that the overall operating load of the server group is too large, and several reference processing methods will be provided to the operation and maintenance personnel. Processing method 1: Increase the number of servers in the server group; Processing method 2: Optimize the algorithm of the load balancing module; Processing method 3: Update the hardware devices in the servers.

[0050] When the server group receives a disaster signal from this data service cloud platform, immediately stop the operations in all servers, and quickly compress and send the data to the mirror backup module

[0051] Such as Figure 2As shown, the mirror backup module is responsible for regularly storing some data in other data service cloud platforms during daily operations. Multiple mirror backup modules, as a whole, completely preserve all the data of all other platforms. A single mirror backup module regularly mirrors some data in the server group of other data service cloud platforms, formats the copied data every preset time, and then re-accesses other data service cloud platforms to mirror the data in each server again. When the mirror backup module receives a disaster signal, it immediately copies the data of the server group of the data service cloud platform where the disaster signal originates, and includes the servers with operating parameters less than the operating parameter threshold into the mirror backup module.

[0052] The disaster recovery module is responsible for sending disaster signals to the mirror backup module during power outages, fires, floods, and other natural disasters. The disaster recovery module includes a power outage sensor, a smoke sensor, and a seismic sensor. When the data obtained by the power outage sensor, the smoke sensor, or the seismic sensor exceeds the threshold, it sends a disaster signal to the mirror backup module of the server group and all other data service cloud platforms, and issues a warning message to the operation and maintenance management personnel through the display screen.

Claims

1. A comprehensive management system for an Internet-based data service cloud platform, including a load balancing module, a mirror backup module, a visualization operation and maintenance module, a data cleaning and classification module, a cache server, a server group, and a disaster recovery module, characterized in that: The load balancing module distributes sub-data streams to each server in the server group through the smooth weighted round-robin algorithm, combines the analysis of the load conditions of the servers to distribute the sub-data streams, and at the same time dynamically adjusts the weights by simulating the adverse effects of the influx of sub-data streams on the load of a single server through an algorithm; The mirror backup module is used to periodically store part of the data in other data service cloud platforms, deletes the stored backup data every preset time, and starts the next backup cycle. Multiple mirror backup modules, as a whole, completely store all the data of all other platforms; it is also used to analyze the source of the disaster signal and preferentially back up the data at the source after receiving the disaster signal; The visualization operation and maintenance module performs data visualization processing on the operating status of the servers. Specifically, it converts the load conditions into a load parameter spectrogram through the fast Fourier transform and compares the server load condition spectrogram with the standard server load coefficient spectrogram. During the comparison process, key data is captured in the image and the key data is fused and comprehensively analyzed to analyze the operating conditions of the server group and the cache server, and the analysis results are fed back to the operation and maintenance personnel and optimization suggestions are put forward to the operation and maintenance personnel; The data cleaning and classification module cleans and classifies the data streams input into the cloud platform. Before the cleaning starts, the original data is backed up, and then the data stream is split into more than two sub-data streams. The documents and annotations in each sub-data stream are read to determine the information type and data type of the sub-data stream; Obtain a preset information type factor according to the information type, and obtain a preset data type factor according to the data type; Calculate the cache factor of each sub-data stream by calculating the information type factor and the data type factor, and based on the comparison result of the cache factor with the preset threshold, distribute each sub-data stream to the cache server and the load balancing module; When the cache factor of a certain sub-data stream is greater than or equal to the preset cache factor threshold, output the sub-data stream to the cache server; when the cache factor of a certain sub-data stream is less than the preset cache factor threshold, output the sub-data stream to the load balancing module; The cache server stores the data streams with more access times, uses a disk with faster read and write speeds, and exchanges data with the server group; The server group is a data processing center composed of multiple servers. Each server in the server group jointly undertakes the data processing tasks through the task distribution of the load balancing module; The disaster recovery module issues a disaster signal when an unexpected situation threatens data security, indicating that the data service cloud platform performs data transfer.

2. The integrated management system of a data service cloud platform based on the Internet according to claim 1, characterized in that, The load balancing module distributes sub-data streams to each server in the server group through the smooth weighted round-robin algorithm. The specific method is as follows: In the first step, the load balancing module sends a heartbeat packet signal to all servers every preset time. After receiving the heartbeat packet signal, the server reads the CPU occupancy rate and the disk read rate; calculates the load weight and sends the load weight to the load balancing module; In the second step, the load balancing module calculates all load weights, and sends the first sub-data stream to the server with the largest load weight. When the largest load weight comes from multiple servers, the first sub-data stream is sent to the server with the smallest number among them, and the load weight is marked. In the third step, all load weights are adjusted. The adjustment process is as follows: First, the marked load weight is decreased by calculation and then the mark is removed; then, the values of other unmarked load weights are increased by calculation. In the fourth step, the load balancing module sends the second sub-data stream to the server with the largest load weight. When the largest load weight comes from multiple servers, the second sub-data stream is sent to the server with the smallest number among them, the load weight of the server is marked, and the third step is returned.

3. The integrated management system of a data service cloud platform based on the Internet according to claim 1, characterized in that, The data backup method of the mirror backup module is specifically as follows: A single mirror backup module periodically performs mirror copying on part of the data in the server group in other data service cloud platforms, formats the copied data every preset time, and re-accesses other data service cloud platforms to perform mirror copying on the data in each server again. When the mirror backup module receives a disaster signal, it immediately copies the data of the server group of the data service cloud platform where the disaster signal comes from, and incorporates the servers with operating parameters less than the preset threshold in the server group into the mirror backup module.

4. A comprehensive management system for an Internet-based data service cloud platform according to claim 1, characterized in that The visualization operation and maintenance module performs data visualization processing on the server operation status, specifically as follows: Record the time number, the CPU occupancy rate of the server group, the disk read rate and the memory occupancy rate, the CPU occupancy rate of the cache server, the disk read rate and the memory occupancy rate, and calculate the operating parameters of the server and the operating parameters of the cache server. Establish a two-dimensional rectangular coordinate system with the time number as the abscissa and the server operating parameters as the ordinate, and obtain a two-dimensional discrete random variable distribution diagram of the server operating parameters changing with the time number for all points. Connect all points with a smooth curve, and then fit the smooth curve with an elementary function image to obtain a change function image of the server group's operating parameters changing with the time label. Similarly, obtain a change function image of the cache server's operating parameters changing with the time label; use the fast Fourier transform to transform the image from the time domain to the frequency domain to obtain the server operating parameter spectrum diagram and the cache server operating parameter spectrum diagram, and compare them with the standard server operating parameter spectrum diagram; calculate the server abnormality parameter and the cache server abnormality parameter from the comparison result, and perform data fusion on the parameters to obtain the comprehensive abnormality parameter. Compare the server operating parameter spectrum diagram with the preset standard server operating parameter spectrum diagram. Extract the spectral feature data from the server operation parameter spectrogram, including the total number of peak points, the sequential numbers of peak points, the operation parameters corresponding to peak points, the frequencies corresponding to peak points, the total number of valley points, the sequential numbers of valley points, the operation parameters corresponding to valley points, and the frequencies corresponding to valley points; extract the spectral feature data from the preset standard server operation parameter spectrogram, including the total number of peak points, the sequential numbers of peak points, the operation parameters corresponding to peak points, the frequencies corresponding to peak points, the total number of valley points, the sequential numbers of valley points, the operation parameters corresponding to valley points, and the frequencies corresponding to valley points; Obtain the server anomaly parameter through the calculation of the spectral feature data of the server operation parameter spectrogram and the preset standard server operation parameter spectrogram; Compare the cache server operation parameter spectrogram with the preset standard cache server operation parameter spectrogram; Extract the spectral feature data from the cache server operation parameter spectrogram, including the total number of peak points, the sequential numbers of peak points, the operation parameters corresponding to peak points, the frequencies corresponding to peak points, the total number of valley points, the sequential numbers of valley points, the operation parameters corresponding to valley points, and the frequencies corresponding to valley points; extract the spectral feature data from the preset standard cache server operation parameter spectrogram, including the total number of peak points, the sequential numbers of peak points, the operation parameters corresponding to peak points, the frequencies corresponding to peak points, the total number of valley points, the sequential numbers of valley points, the operation parameters corresponding to valley points, and the frequencies corresponding to valley points; Obtain the cache server anomaly parameter through the calculation of the spectral feature data of the cache server operation parameter spectrogram and the preset standard cache server operation parameter spectrogram; Obtain the comprehensive anomaly parameter through the calculation of the server anomaly parameter and the cache server anomaly parameter.

5. The integrated management system of a data service cloud platform based on the Internet according to claim 1, characterized in that, The cache server and the server group process the data stream and record the load data, specifically: After receiving the data stream, the cache server records the time number, CPU occupancy rate, disk read rate, and memory occupancy rate at every preset time interval and sends them to the visualization operation and maintenance module; After receiving the data stream, the server group records the time number, CPU occupancy rate, disk read rate, and memory occupancy rate at every preset time interval and sends them to the visualization operation and maintenance module.

6. The integrated management system of a data service cloud platform based on the Internet according to claim 1, characterized in that, Perform data transfer when an emergency threatens data security, specifically: The disaster recovery module includes a power-off sensor, a smoke sensor, and a seismic sensor. When the data obtained by the power-off sensor or the smoke sensor or the seismic sensor exceeds the threshold, it sends a disaster signal to the server group and the mirror backup modules of all other data service cloud platforms, and issues a warning message to the operation and maintenance management personnel through the display screen; When the mirror backup module receives the disaster signal, it immediately copies the data of the server group of the data service cloud platform from which the disaster signal originates, and includes the servers in the server group with operation parameters less than the operation parameter threshold in the mirror backup module.

Images