A monitoring method based on active-active switching control platform

Abstract

The present invention discloses a monitoring method based on an active-active switching control platform, and the method includes data collection, data preprocessing, data dimension reduction, clustering processing and active-active switching control platform monitoring. The present invention belongs to the field of platform monitoring technology, and specifically refers to a monitoring method based on an active-active switching control platform. This scheme maintains the local structure of the data based on the nearest neighbor matrix and basis function, reduces the reconstruction error by optimizing the cost function and applying the Lagrangian function; performs dimension reduction processing on the data through the optimal solution and nonlinear mapping function to reduce the amount of monitoring calculations; uses Delaunay triangulation to construct a triangular mesh, uses the distance between the data points and the origin to arrange in ascending order, and determines the merging conditions based on the similarity of the normal vector, which helps to maintain the accuracy and reliability of the clustering results; by searching and automatically merging adjacent triangles, it can automatically identify and assign all triangles to discontinuous surfaces; thereby making the platform monitoring results more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of platform monitoring, and in particular to a monitoring method based on a dual-active switching management and control platform. Background Art

[0002] The monitoring method based on the active-active switching control platform generally refers to a technology and strategy for monitoring and managing active-active systems. The active-active switching control platform is a high-availability architecture that allows the system to share loads and requests between multiple active nodes in real time to ensure that the service is still available even if one of the nodes fails or requires maintenance. However, the general control platform monitoring method has problems such as excessive data dimensions leading to excessive calculations, inability to maintain the local structure of the data when reducing the dimension of the platform monitoring data, and excessive data reconstruction errors, resulting in low data feature retention; the general control platform monitoring method has problems such as too much data when clustering the platform monitoring data, resulting in slow clustering convergence and high time consumption, and poor adaptability of clustering processing when the platform monitoring data is unevenly distributed. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a monitoring method based on a dual-active switching control platform. In view of the problems that the general control platform monitoring method has too large data dimensions, resulting in too much calculation, and the local structure of the data cannot be maintained when the platform monitoring data is reduced in dimension, and the data reconstruction error is too large, resulting in low data feature retention, this scheme is based on local nonlinear mapping functions to capture the complex relationship of nonlinear structures in the data, based on the nearest neighbor matrix and basis functions to maintain the local structure of the data, and by optimizing the cost function and applying the Lagrangian function to reduce the reconstruction error, to ensure the rationality of the weight coefficient and the stability of the data; the data is reduced in dimension through the optimal solution and nonlinear mapping function to reduce the monitoring calculation amount; in view of the general The monitoring method of the control platform has the problem that when clustering the platform monitoring data, the amount of data is too large, resulting in slow clustering convergence and a large amount of time consumption. When the platform monitoring data is unevenly distributed, the adaptability of clustering processing is poor. This solution uses Delaunay triangulation to construct a triangular mesh, uses the distance between the data points and the origin to arrange them in ascending order, and determines the merging conditions based on the similarity of the normal vector, which helps to maintain the accuracy and reliability of the clustering results and can effectively process large-scale data sets; by searching and automatically merging adjacent triangles, it can automatically identify and assign all triangles to discontinuous surfaces, thereby optimizing data; it performs well when it needs to process irregular shapes and non-uniformly distributed data; and thus makes the platform monitoring results more accurate.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a monitoring method based on a dual-active switching management and control platform, the method comprising the following steps:

[0005] Step S1: data collection;

[0006] Step S2: data preprocessing;

[0007] Step S3: data dimension reduction;

[0008] Step S4: clustering processing;

[0009] Step S5: Active-active switching control platform monitoring.

[0010] Furthermore, in step S1, the data collection is to collect historical platform monitoring data and real-time platform monitoring data; the historical platform monitoring data and real-time platform monitoring data both include system performance data, application performance data, active-active status monitoring data, user behavior data and environment monitoring data; the system performance data includes CPU usage, memory usage and network bandwidth usage; the application performance data includes response time, throughput and error rate; the historical platform monitoring data also includes platform detection status; the platform detection status includes normal operation and abnormal operation; the platform detection status is used as a data label; the data label does not participate in subsequent clustering operations, and is only selected and used as a cluster label.

[0011] Furthermore, in step S2, the data preprocessing is to perform data cleaning, data conversion and standardization on the collected historical platform monitoring data and real-time platform monitoring data; the data cleaning is to process missing values, duplicate values ​​and outliers; the data conversion is to convert the data into vector form; the standardization is to standardize the data based on the maximum and minimum normalization method.

[0012] Furthermore, in step S3, the data dimension reduction is to perform dimension reduction processing on the preprocessed data; specifically, the following steps are included:

[0013] Step S31: Construct a local nonlinear mapping function using an independent basis function sequence , the local nonlinear region is expressed as:

[0014] ;

[0015] Where g(·) is the nonlinear mapping function; is the data point x i The nearest neighbor matrix of is the kth independent basis function; m is the total number of basis functions; is the weight coefficient of the kth basis function, expressed as a row vector;

[0016] Step S32: Cost function, converting data point x i The cost function is expressed as:

[0017] ;

[0018] Where K is the total number of data points; w ij is the weight coefficient of the data point, indicating that the data point x i The weight between the jth nearest neighbor data point; X is the data point matrix; is the jth nearest neighbor data point of the ith data point; W is the weight vector, which contains the weight coefficients w of all data points ij ;

[0019] Step S33: Optimization problem, by solving the following optimization problem to obtain the weight coefficient {a i}:

[0020] ;

[0021] In the formula, represents the reconstruction cost function; a i is the weight coefficient of the ith basis function; E is a vector of all 1s; T is the transpose; a1 is the weight coefficient of the first basis function;

[0022] Step S34: Lagrange multiplier method, using Lagrange multiplier method to convert the optimization problem into an unconstrained optimization problem, which is expressed as follows:

[0023] ;

[0024] Where L(·) is the Lagrangian function, γ is the Lagrangian multiplier; a2 is the weight coefficient of the second basis function; and are the kth and rth basis functions respectively; a r is the weight coefficient of the rth basis function; (·,·) is the inner product operation;

[0025] Step S35: Find the optimal solution, calculate the gradient and set it to zero to find the optimal solution {a i}, the formula used is as follows:

[0026] ;

[0027] In the formula, is the partial derivative of the Lagrangian function with respect to the jth weight coefficient; is the jth basis function; is the Kronecker symbol. When j=1, , otherwise 0;

[0028] Step S36: Data dimensionality reduction, mapping the data points based on the weight coefficient and the nonlinear mapping function to obtain three-dimensional data, and then obtain a data set after dimensionality reduction processing; the formula used is as follows:

[0029] ;

[0030] In the formula, y i is the data point x i The result after dimensionality reduction.

[0031] Furthermore, in step S4, the clustering process specifically includes the following steps:

[0032] Step S41: construct a mesh, arrange the data points in ascending order according to the distance between the data points and the origin, update the vertex index of the triangle, construct a triangular mesh based on Delaunay triangulation, and obtain a triangle set;

[0033] Step S42: Search, select a triangle from the triangle set as a seed triangle, define its search radius, and use it as the starting point for searching; find triangles that share common edges with the seed triangle; test whether these adjacent triangles meet the merging conditions based on normal vector similarity; the triangles that meet the conditions will be added to the current area and become new seed triangles, and continue the next round of searching and checking; repeat the selection of new seed triangles and the search and merging until all triangles are assigned to discontinuous faces; obtain a continuous area set, each area consisting of multiple adjacent triangles;

[0034] Step S43: fuzzy clustering, performing fuzzy clustering operations on data points in different regions respectively; and obtaining clustering results.

[0035] Further, in step S5, the active-active switching management and control platform monitoring is based on the clustering result, and the label type containing the largest number of historical data in the cluster is used as the cluster label; the cluster label of the real-time platform monitoring data is used as the platform monitoring result corresponding to the data.

[0036] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0037] (1) In view of the problems that the general monitoring method of the control platform has the problem of too large data dimension resulting in excessive calculation, the local structure of the data cannot be maintained when the platform monitoring data is reduced in dimension, and the data reconstruction error is too large, resulting in low data feature retention. This scheme is based on the local nonlinear mapping function to capture the complex relationship of the nonlinear structure in the data, based on the nearest neighbor matrix and basis function to maintain the local structure of the data, and by optimizing the cost function and applying the Lagrangian function to reduce the reconstruction error, ensure the rationality of the weight coefficient and the stability of the data; reduce the dimension of the data through the optimal solution and nonlinear mapping function to reduce the amount of monitoring calculation.

[0038] (2) In view of the problems that the general control platform monitoring method has when clustering the platform monitoring data, the amount of data is too large, resulting in slow clustering convergence and a long time consumption. When the platform monitoring data is unevenly distributed, the adaptability of clustering processing is poor. This scheme uses Delaunay triangulation to construct a triangular mesh, uses the distance between the data points and the origin to arrange them in ascending order, and determines the merging conditions based on the similarity of normal vectors. This helps to maintain the accuracy and reliability of the clustering results and can effectively process large-scale data sets. By searching and automatically merging adjacent triangles, it can automatically identify and assign all triangles to discontinuous surfaces, thereby optimizing data. It performs well when it needs to process irregular shapes and non-uniformly distributed data, thereby making the platform monitoring results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flow chart of a monitoring method based on a dual-active switching control platform provided by the present invention;

[0040] Figure 2 is a schematic flow chart of step S3;

[0041] Figure 3 It is a schematic diagram of the process of step S4.

[0042] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0044] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0045] Example 1, see Figure 1 The present invention provides a monitoring method based on a dual-active switching management and control platform, the method comprising the following steps:

[0046] Step S1: Data collection, collecting historical platform monitoring data and real-time platform monitoring data;

[0047] Step S2: Data preprocessing, data cleaning, data conversion and standardization of the collected historical platform monitoring data and real-time platform monitoring data;

[0048] Step S3: Data dimensionality reduction, capturing the complex relationship of nonlinear structure in the data based on the local nonlinear mapping function, maintaining the local structure of the data based on the nearest neighbor matrix and basis function, reducing the reconstruction error by optimizing the cost function and applying the Lagrangian function; reducing the dimensionality of the data through the optimal solution and nonlinear mapping function;

[0049] Step S4: Clustering processing, using Delaunay triangulation to construct a triangular mesh, using the distance between the data points and the origin to arrange them in ascending order, based on the similarity of the normal vectors to determine the merging conditions, to achieve data optimization, and finally perform fuzzy clustering processing on the data;

[0050] Step S5: Active-active switching control platform monitoring.

[0051] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, both the historical platform monitoring data and the real-time platform monitoring data include system performance data, application performance data, active-active status monitoring data, user behavior data and environment monitoring data; the system performance data includes CPU usage, memory usage and network bandwidth usage; the application performance data includes response time, throughput and error rate; the historical platform monitoring data also includes platform detection status; the platform detection status includes normal operation and abnormal operation; the platform detection status is used as a data label; the data label does not participate in subsequent clustering operations and is selected and used as a cluster label.

[0052] Example 3, see Figure 1 This embodiment is based on the above embodiment. In step S2, data cleaning is to process missing values, duplicate values ​​and outliers; data conversion is to convert data into vector form; and standardization is to standardize data based on the maximum and minimum normalization method.

[0053] Example 4, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. In step S3, data dimension reduction is to perform dimension reduction processing on the preprocessed data. Specifically, the following steps are included:

[0054] Step S31: Construct a local nonlinear mapping function using an independent basis function sequence , the local nonlinear region is expressed as:

[0055] ;

[0056] Where g(·) is the nonlinear mapping function; is the data point x i The nearest neighbor matrix of is the kth independent basis function; m is the total number of basis functions; is the weight coefficient of the kth basis function, expressed as a row vector;

[0057] Step S32: Cost function, converting data point x i The cost function is expressed as:

[0058] ;

[0059] Where K is the total number of data points; w ij is the weight coefficient of the data point, indicating that the data point x i The weight between the jth nearest neighbor data point; X is the data point matrix; is the jth nearest neighbor data point of the ith data point; W is the weight vector, which contains the weight coefficients w of all data points ij ;

[0060] Step S33: Optimization problem, by solving the following optimization problem to obtain the weight coefficient {a i}:

[0061] ;

[0062] In the formula, represents the reconstruction cost function; a i is the weight coefficient of the ith basis function; E is a vector of all 1s; T is the transpose; a1 is the weight coefficient of the first basis function;

[0063] Step S34: Lagrange multiplier method, using Lagrange multiplier method to convert the optimization problem into an unconstrained optimization problem, which is expressed as follows:

[0064] ;

[0065] Where L(·) is the Lagrangian function, γ is the Lagrangian multiplier; a2 is the weight coefficient of the second basis function; and are the kth and rth basis functions respectively; a r is the weight coefficient of the rth basis function; (·,·) is the inner product operation;

[0066] Step S35: Find the optimal solution, calculate the gradient and set it to zero to find the optimal solution {a i}, the formula used is as follows:

[0067] ;

[0068] In the formula, is the partial derivative of the Lagrangian function with respect to the jth weight coefficient; is the jth basis function; is the Kronecker symbol. When j=1, , otherwise 0;

[0069] Step S36: Data dimensionality reduction, mapping the data points based on the weight coefficient and the nonlinear mapping function to obtain three-dimensional data, and then obtain a data set after dimensionality reduction processing; the formula used is as follows:

[0070] ;

[0071] In the formula, y i is the data point x i The result after dimensionality reduction.

[0072] By performing the above operations, the general management and control platform monitoring method has the problem that the data dimension is too large, resulting in too much calculation, the local structure of the data cannot be maintained when the platform monitoring data is reduced in dimension, and the data reconstruction error is too large, resulting in low data feature retention. This solution is based on local nonlinear mapping functions to capture the complex relationship of nonlinear structures in the data, based on the nearest neighbor matrix and basis function to maintain the local structure of the data, and by optimizing the cost function and applying the Lagrangian function to reduce the reconstruction error, ensuring the rationality of the weight coefficient and the stability of the data; the data is reduced in dimension through the optimal solution and nonlinear mapping function to reduce the amount of monitoring calculations.

[0073] Example 5, see Figure 1 and Figure 3 This embodiment is based on the above embodiment. In step S4, the clustering process specifically includes the following steps:

[0074] Step S41: construct a mesh, arrange the data points in ascending order according to the distance between the data points and the origin, update the vertex index of the triangle, construct a triangular mesh based on Delaunay triangulation, and obtain a triangle set;

[0075] Step S42: Search, select a triangle from the triangle set as a seed triangle, define its search radius, and use it as the starting point for searching; find triangles that share common edges with the seed triangle; test whether these adjacent triangles meet the merging conditions based on normal vector similarity; the triangles that meet the conditions will be added to the current area and become new seed triangles, and continue the next round of searching and checking; repeat the selection of new seed triangles and the search and merging until all triangles are assigned to discontinuous faces; obtain a continuous area set, each area consisting of multiple adjacent triangles;

[0076] Step S43: fuzzy clustering, performing fuzzy clustering operations on data points in different regions respectively; and obtaining clustering results.

[0077] By performing the above operations, the general management and control platform monitoring method has the problem that when clustering platform monitoring data, the amount of data is too large, resulting in slow clustering convergence and high time consumption. When the platform monitoring data is unevenly distributed, the adaptability of clustering processing is poor. This solution uses Delaunay triangulation to construct a triangular mesh, uses the distance between the data points and the origin to arrange them in ascending order, and determines the merging conditions based on the similarity of normal vectors, which helps to maintain the accuracy and reliability of the clustering results and can effectively process large-scale data sets; by searching and automatically merging adjacent triangles, it can automatically identify and assign all triangles to discontinuous surfaces, thereby optimizing data; it performs well when it needs to process irregular shapes and non-uniformly distributed data, thereby making the platform monitoring results more accurate.

[0078] Example 6, see Figure 1 This embodiment is based on the above embodiment. In step S5, the active-active switching management and control platform monitoring is based on the clustering result, and the label type containing the largest number of historical data in the cluster is used as the cluster label; the cluster label of the real-time platform monitoring data is used as the platform monitoring result corresponding to the data.

[0079] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0080] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0081] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A monitoring method based on a dual-active switching management and control platform, characterized in that: The method comprises the following steps: Step S1: data collection; Step S2: data preprocessing; Step S3: Data dimensionality reduction, capturing the complex relationship of nonlinear structure in the data based on the local nonlinear mapping function, maintaining the local structure of the data based on the nearest neighbor matrix and basis function, reducing the reconstruction error by optimizing the cost function and applying the Lagrangian function; reducing the dimensionality of the data through the optimal solution and nonlinear mapping function; Step S4: Clustering processing, using Delaunay triangulation to construct a triangular mesh, using the distance between the data points and the origin to arrange them in ascending order, based on the similarity of the normal vectors to determine the merging conditions, to achieve data optimization, and finally perform fuzzy clustering processing on the data; Step S5: Active-active switching control platform monitoring; In step S1, the data collection is to collect historical platform monitoring data and real-time platform monitoring data; the historical platform monitoring data and real-time platform monitoring data both include system performance data, application performance data, active-active status monitoring data, user behavior data and environment monitoring data; the system performance data includes CPU usage, memory usage and network bandwidth usage; the application performance data includes response time, throughput and error rate; the historical platform monitoring data also includes platform detection status; the platform detection status includes normal operation and abnormal operation; the platform detection status is used as a data label; the data label does not participate in subsequent clustering operations, and is selected and used as a cluster label; In step S2, the data preprocessing is to clean, convert and standardize the collected historical platform monitoring data and real-time platform monitoring data; the data cleaning is to process missing values, duplicate values ​​and outliers; the data conversion is to convert the data into vector form; the standardization is to standardize the data based on the maximum and minimum normalization method.

2. According to a monitoring method based on a dual-active switching control platform according to claim 1, it is characterized in that: In step S4, the clustering process specifically includes the following steps: Step S41: construct a mesh, arrange the data points in ascending order according to the distance between the data points and the origin, update the vertex index of the triangle, construct a triangular mesh based on Delaunay triangulation, and obtain a triangle set; Step S42: Search, select a triangle from the triangle set as a seed triangle, define its search radius, and use it as the starting point for searching; find triangles that share common edges with the seed triangle; test whether these adjacent triangles meet the merging conditions based on normal vector similarity; the triangles that meet the conditions will be added to the current area and become new seed triangles, and continue the next round of searching and checking; repeat the selection of new seed triangles and the search and merging until all triangles are assigned to discontinuous faces; obtain a continuous area set, each area consisting of multiple adjacent triangles; Step S43: fuzzy clustering, performing fuzzy clustering operations on data points in different regions respectively; and obtaining clustering results.

3. According to a monitoring method based on a dual-active switching control platform according to claim 1, it is characterized in that: In step S3, the data dimension reduction is to perform dimension reduction processing on the preprocessed data; specifically, the following steps are included: Step S31: Construct a local nonlinear mapping function using an independent basis function sequence , the local nonlinear region is expressed as: ； Where g(·) is the nonlinear mapping function; is the data point x i The nearest neighbor matrix of is the kth independent basis function; m is the total number of basis functions; is the weight coefficient of the kth basis function, expressed as a row vector; Step S32: Cost function, transforming data point x i The cost function is expressed as: ； Where K is the total number of data points; w ij is the weight coefficient of the data point, indicating that the data point x i The weight between the jth nearest neighbor data point; X is the data point matrix; is the jth nearest neighbor data point of the ith data point; W is the weight vector, which contains the weight coefficients w of all data points ij ; Step S33: Optimization problem, by solving the following optimization problem to obtain the weight coefficient {a i }: ； In the formula, represents the reconstruction cost function; a i is the weight coefficient of the ith basis function; E is a vector of all 1s; T is the transpose; a1 is the weight coefficient of the first basis function; Step S34: Lagrange multiplier method, using Lagrange multiplier method to convert the optimization problem into an unconstrained optimization problem, which is expressed as follows: ； Where L(·) is the Lagrangian function, γ is the Lagrangian multiplier; a2 is the weight coefficient of the second basis function; and are the kth and rth basis functions respectively; a r is the weight coefficient of the rth basis function; (·,·) is the inner product operation; Step S35: Find the optimal solution, calculate the gradient and set it to zero to find the optimal solution {a i }, the formula used is as follows: ； In the formula, is the partial derivative of the Lagrangian function with respect to the jth weight coefficient; is the jth basis function; is the Kronecker symbol. When j=1, , otherwise 0; Step S36: Data dimensionality reduction, mapping the data points based on the weight coefficient and the nonlinear mapping function to obtain three-dimensional data, and then obtain a data set after dimensionality reduction processing; the formula used is as follows: ； In the formula, y i is the data point x i The result after dimensionality reduction.

4. According to a monitoring method based on a dual-active switching control platform according to claim 1, it is characterized in that: In step S5, the active-active switching control platform monitors the clustering result and uses the tag type containing the largest number of historical data in the cluster as the cluster tag; The cluster labels of real-time platform monitoring data are used as the platform monitoring results corresponding to the data.

Images