A productivity middle platform micro-service isolated running method, medium and system

Abstract

The application provides a productivity middle platform micro-service isolated operation method, medium and system, belongs to the electric digital data processing technical field, the application proposes a productivity middle platform micro-service isolated operation method, realizes micro-service resource accurate scheduling through multidimensional dynamic analysis. The method first collects micro-service operation indexes, including processor occupancy, memory occupancy, network throughput and the like, and generates micro-service priority indexes; then, the operation indexes are decomposed and operated to construct a benchmark and a dynamic resource matrix; further, transient and persistent competition indexes are calculated to establish a micro-service call graph; a resource allocation equation set is constructed to determine a resource allocation scheme; an optimized dichotomy algorithm is used to determine a resource use threshold value; finally, resource unit loads are monitored in real time, when the threshold value is exceeded, migration cost is calculated and micro-service isolation operation is performed, solving the technical problem that micro-service resource allocation and isolation mechanism cannot be accurately and dynamically adjusted in the prior art.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of digital data processing, and particularly relates to a productivity middle platform micro-service isolated operation method, medium and system. BACKGROUND

[0002] As a key technology of modern distributed systems, micro-service architecture has been widely applied in cloud computing, Internet services, financial technology and other fields. Traditional micro-service resource management mainly relies on static resource allocation mode, and resource scheduling is performed through manual configuration or simple load balancing algorithm. This method has significant limitations in resource utilization and service performance, and cannot effectively cope with dynamic changes in business load and complex service dependency.

[0003] Traditional micro-service resource allocation techniques mainly include static resource reservation, average load-based balanced allocation, and simple resource threshold monitoring methods. These methods are difficult to accurately capture the real-time running characteristics of micro-services, and resource allocation often has hysteresis and inaccuracy. Especially in high-concurrency and variable business scenarios, static resource allocation mode will lead to low resource utilization efficiency, service performance fluctuations and other problems.

[0004] The core technical challenge faced by the prior art is the lack of a comprehensive technical solution that can dynamically, accurately and intelligently allocate and isolate micro-service resources. Traditional methods cannot effectively balance service performance, resource utilization and system stability, and are difficult to cope with dynamic resource scheduling needs in complex micro-service environments. That is, there is a technical problem in the prior art that the micro-service resource allocation and isolation mechanism cannot be accurately and dynamically adjusted. SUMMARY

[0005] Therefore, the present application provides a productivity middle platform micro-service isolated operation method, medium and system, which can solve the technical problem that the micro-service resource allocation and isolation mechanism cannot be accurately and dynamically adjusted in the prior art.

[0006] The present application is implemented as follows: In a first aspect, the present application provides a productivity middle platform micro-service isolated operation method for collecting micro-service running index data, performing decomposition operation on the micro-service running index data to obtain basic load value and dynamic load value, constructing a benchmark resource matrix and a dynamic resource matrix, calculating a competition index based on the benchmark resource matrix and the dynamic resource matrix, constructing a micro-service call graph, establishing a micro-service correlation matrix, converting micro-service grouping into a multi-dimensional constrained knapsack problem, and solving a resource allocation scheme using a resource allocation equation group including a utility calculation equation, a correlation calculation equation, a load balancing equation, and a migration cost equation. Based on the resource balance constraint index, determine the resource usage threshold value and divide the resource unit, monitor the resource unit load index and perform micro-service isolation operation.

[0007] The micro-service running index data includes a processor occupancy rate, a memory occupancy rate, a network throughput, a disk read-write rate, a service response delay, a service call frequency, a data exchange volume, and environmental parameter data, and a micro-service priority index is generated through the micro-service running index data.

[0008] The competition index includes a transient competition index and a persistent competition index, the transient competition index is calculated based on the dynamic resource matrix, the persistent competition index is calculated based on the benchmark resource matrix, and a resource request vector is generated in combination with the transient competition index and the persistent competition index.

[0009] The correlation strength index between the micro-services is analyzed by using path optimization operation, and the micro-service correlation matrix is established.

[0010] The resource allocation equation set includes an utility calculation equation, a correlation degree calculation equation, a load balancing equation, and a migration cost equation.

[0011] The input of the utility calculation equation includes the resource request vector, the resource unit capacity data, and the micro-service priority index, and the output is a micro-service utility allocation matrix; the input of the correlation degree calculation equation includes the service call frequency, the data exchange volume, the correlation strength index, and the service response delay, and the output is the micro-service correlation matrix.

[0012] The input of the load balancing equation includes the resource unit load index, the load change record, the micro-service priority index, and the resource usage threshold value, and the output is the resource balancing constraint index; the input of the migration cost equation includes the state migration data volume, the network channel occupancy rate, the migration operation time, and the service influence degree, and the output is a migration cost index.

[0013] When the resource unit load index exceeds the resource usage threshold value, a migration execution path is determined based on the state migration data volume, the network channel occupancy rate, the migration operation time, and the service influence degree, and a micro-service isolation operation is completed.

[0014] The second aspect of the application provides a computer readable storage medium, the computer readable storage medium stores program instructions, the program instructions are used to execute the productivity middle platform micro-service isolation running method when running in the computer.

[0015] The third aspect of the present application provides a productivity middle platform microservice isolated running system, comprising the computer readable storage medium described above, the system is any one of a computer, a server, and a single chip microcomputer, the computer readable storage medium is arranged in the system, and the system is provided with a microprocessor for executing program instructions stored in the computer readable storage medium.

[0016] Compared with the prior art, the productivity middle platform microservice isolated running method, medium and system provided by the present application, the present application proposes a microservice resource isolated running method based on multi-dimensional dynamic analysis, which revolutionizes the microservice resource management technology by constructing an intelligent and accurate resource allocation model. The method breaks through the limitations of the traditional static resource allocation mode and introduces an intelligent resource scheduling mechanism based on load characteristics, service correlation and dynamic competition.

[0017] The core innovation of the method is to establish a multi-level, self-adaptive resource analysis and allocation system. Through multi-dimensional data collection, complex mathematical modeling and dynamic optimization algorithm, the method realizes fine and intelligent adjustment of microservice resource allocation. The method not only accurately captures the running characteristics of microservices, but also dynamically evaluates the correlation between services and resource competition.

[0018] Compared with the traditional technology, the technical scheme of the present application fundamentally solves the problems of dynamic, accuracy and adaptability of microservice resource allocation. By introducing advanced mathematical models and intelligent algorithms, the resource utilization efficiency, service performance stability and overall operation quality of the distributed system are significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The flowchart of the method of the present application.

[0020] Figure 2 The multi-dimensional performance index radar chart of the financial management microservice in Example 2.

[0021] Figure 3 The load decomposition and priority analysis chart in Example 2.

[0022] Figure 4 The competition index and correlation degree analysis chart in Example 2.

[0023] Figure 5 The utility function surface chart in Example 2.

[0024] Figure 6 The migration decision threshold analysis chart in Example 2. DETAILED DESCRIPTION

[0025] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0026] As shown in Figure 1 is a flow chart of a productivity middle platform microservice isolated operation method provided by the first aspect of the present application, and the method comprises the following steps:

[0027] S10, collect microservice running index data, including processor occupancy, memory occupancy, network throughput, disk read / write rate, service response delay, service call frequency, data exchange volume, and environment parameter data, to generate microservice priority index;

[0028] S20, perform decomposition operation on the microservice running index data to obtain a basic load value and a dynamic load value, and record the dynamic load value to form a load change record;

[0029] S30, establish a benchmark resource matrix based on the basic load value, establish a dynamic resource matrix based on the dynamic load value, and generate resource unit capacity data according to the benchmark resource matrix;

[0030] S40, calculate an instantaneous competition index based on the dynamic resource matrix, calculate a persistent competition index based on the benchmark resource matrix, and generate a resource request vector by combining the instantaneous competition index and the persistent competition index;

[0031] S50, construct a microservice call graph, analyze the correlation strength index between microservices by using path optimization operation, and establish a microservice correlation matrix;

[0032] S60, based on the microservice correlation matrix, the instantaneous competition index and the persistent competition index, construct a resource allocation equation set to solve a resource allocation scheme, and obtain a resource balance constraint index;

[0033] S70, determine a resource usage threshold value based on the resource balance constraint index by using an optimized bisection algorithm, and divide resource units according to the resource usage threshold value;

[0034] S80, monitor the resource unit load index, and when the resource unit load index exceeds the resource usage threshold value, calculate state migration data volume, network channel occupancy, migration operation time and service influence degree, determine a migration execution path, and complete microservice isolation operation.

[0035] The resource allocation equation set comprises an utility calculation equation, a correlation degree calculation equation, a load balancing equation and a migration cost equation;

[0036] The utility calculation equation is used to calculate the utility index of the microservice in the resource unit, the input includes the resource request vector, the resource unit capacity data, the microservice priority index, and the output is a microservice utility allocation matrix;

[0037] The correlation degree calculation equation is used to evaluate the correlation degree between microservices, the input includes the service call frequency, the data exchange amount, the correlation strength index, and the service response delay, and the output is the microservice correlation matrix;

[0038] The load balancing equation is used to ensure the balance of resource allocation, the input includes the resource unit load index, the load change record, the microservice priority index, and the resource usage threshold, and the output is the resource balance constraint index;

[0039] The migration cost equation is used to calculate the overhead of the migration operation, the input includes the state migration data amount, the network channel occupancy rate, the migration operation time, and the service impact degree, and the output is the migration cost index.

[0040] The specific implementation of step S10 is to collect microservice running index data through an advanced distributed monitoring system. First, a comprehensive index collection framework is built, and a lightweight monitoring agent is deployed in the microservice running environment to capture real-time processor occupancy, memory occupancy, network throughput, disk read / write rate, and other underlying resource indicators. The monitoring agent uses non-intrusive data collection technology to obtain accurate resource usage through operating system interfaces and container monitoring interfaces. For service response delay and service call frequency, a tracking component is implanted in the service call link to record the time, input parameters, output parameters, and other key information of each service call. The data exchange amount is accurately calculated by the network traffic monitoring module. The environmental parameter data is collected by a sensor array deployed in the server room, including temperature, humidity, air pressure, and other physical indicators.

[0041] On the basis of index data collection, the collected raw data needs to be standardized and normalized to eliminate the influence of different index dimensions. Principal component analysis dimension reduction method is used to compress multi-dimensional index data to a lower dimension while retaining the main information features of the data. A calculation model of the microservice priority index is constructed, and a machine learning algorithm is introduced, especially an integrated learning method such as gradient boosting decision tree. The model is trained according to historical running data to adaptively determine the weight of each index. The reference value range of the priority index is between 0 and 1, and the closer to 1, the more important the microservice is. For the influence of environmental parameters, a nonlinear mapping function is designed to convert environmental factors into a correction coefficient of resource performance. This method can comprehensively and accurately characterize the running characteristics and importance of microservices.

[0042] The specific implementation of step S20 is a complex decomposition operation on microservice running index data. Using time series decomposition theory, the total load is divided into trend items, seasonal items and random disturbance items. The extraction of basic load is realized by exponential moving average algorithm, which performs weighted average on historical data, giving higher weight to recent data. The calculation of dynamic load introduces wavelet transform technology, which can effectively separate short-term fluctuation components. The error term introduces a Gaussian white noise model, with a standard deviation controlled within the range of 0.01 to 0.05, ensuring the randomness and smoothness of the decomposition process.

[0043] In the decomposition process, first, the original load data is preprocessed to remove outliers and burrs. The median removal method and box plot method are used to identify and eliminate extreme data points. For time series data, difference transformation is used to eliminate the non-stationarity of the data. The extraction of basic load uses Holt's exponential smoothing method, which can well capture long-term trends. The extraction of dynamic load uses multi-scale decomposition technology of wavelet transform, which accurately separates short-term fluctuation components by selecting appropriate wavelet basis functions. The introduction of error term not only improves the numerical accuracy of decomposition, but also simulates the unmeasurable small disturbances in the actual system. This multi-level, multi-scale decomposition method can maximize the revelation of the internal structural characteristics of microservice load.

[0044] The specific implementation of step S30 is to construct a resource matrix with multi-dimensional semantics. The construction of the benchmark resource matrix uses tensor decomposition theory to map multi-dimensional heterogeneous data to a low-dimensional semantic space. First, the original index data collected is standardized for preprocessing, eliminating the dimensional differences between different indicators. Principal component analysis is used for dimension reduction, retaining the main components with an explanation variance contribution rate of more than 95%. The rows of the matrix represent microservice instances, the columns represent different resource dimensions, and the values of each cell reflect the basic load characteristics of the corresponding microservice on a specific resource.

[0045] The construction of dynamic resource matrix introduces dynamic tensor decomposition method, which can capture the time evolution characteristics of resource load. The CANDECOMP / PARAFAC (CP) decomposition algorithm in tensor decomposition is used to decompose multi-dimensional data into low-rank core tensor and factor matrix. Each element in the matrix not only reflects the current resource load, but also contains the trend information of load change. To enhance the semantic expression ability of the matrix, context-aware feature engineering technology is introduced, which mines the implicit dependency relationship between microservices through correlation analysis. For sparse or missing data, matrix completion algorithm is used for intelligent filling to ensure the integrity and consistency of the matrix. This multi-dimensional semantic resource matrix provides rich semantic information for subsequent resource allocation decisions.

[0046] The specific implementation of step S40 is to calculate the competition index through complex statistical and information theory methods. The calculation of the instantaneous competition index introduces the coefficient of variation and entropy weight theory, which not only considers the absolute load size of the resource, but also focuses on the uniformity of the load distribution. The weight allocation method based on Shannon entropy is used to adaptively determine the importance of each resource dimension. The calculation of the persistent competition index combines long-term sequence analysis and probability density estimation technology, and the kernel density estimation method is used to characterize the distribution characteristics of the basic load.

[0047] In the calculation of the competition index, first, the stationarity of the original data is tested using the extended augmented unit root test (ADF) method. For each resource dimension, the coefficient of variation is calculated to measure the dispersion of the load. The Shannon entropy weight method is introduced, and the weight of each resource dimension is automatically calculated based on the dispersion of the data. The greater the weight, the higher the uncertainty of the dimension. The calculation of the persistent competition index uses Gaussian kernel density estimation, which can smoothly reconstruct the probability density function of the basic load. The selection of the adjustment coefficient is determined through cross-validation and grid search methods to ensure the robustness of the competition index calculation. This multi-dimensional and multi-level competition index calculation method can comprehensively and accurately characterize the resource contention characteristics of microservices.

[0048] The specific implementation of step S50 is to construct a microservice call graph. The call relationship between microservices is abstracted as a directed and weighted graph using complex network analysis theory. First, the detailed metadata of service calls, including call frequency, data exchange volume, and response delay, are collected through a distributed tracing system (such as OpenTelemetry). Community detection algorithms in graph theory are used to identify tightly connected clusters between microservices. The PageRank and betweenness centrality of complex network centrality indicators are introduced to quantify the importance of microservices in the overall system architecture. The correlation strength calculation uses the mutual information theory in information theory, considering the call frequency and data exchange characteristics between services. When constructing the correlation matrix, compression storage technology is used for sparse matrices to improve the computational efficiency of large-scale service networks. Graph embedding technology is used to map high-dimensional service correlation to low-dimensional semantic space, providing topological semantic information for subsequent resource allocation decisions.

[0049] The specific implementation of step S60 is to construct a resource allocation equation set. By using multi-objective optimization theory, the resource allocation problem is converted into a Pareto optimal problem under constraints. The utility calculation equation introduces the concept of utility function in game theory, considering the priority of microservices, resource request and time decay factor. The correlation degree calculation equation uses the mutual information and entropy theory in information theory to quantify the dependency between microservices. The load balancing equation uses the random process theory to introduce the Markov chain model to describe the stationarity and transition probability of load changes. The migration cost equation combines decision theory and risk assessment methods to construct a multi-dimensional weighted cost model. When solving the equation set, the improved particle swarm optimization algorithm and the simulated annealing algorithm are used to find the global optimal solution in the complex constraint space. By introducing the Lagrange dual theory, the complex multi-objective optimization problem is converted into an equivalent dual problem, improving the numerical stability and computational efficiency of the solution.

[0050] The specific implementation of step S70 is to determine the resource usage threshold value using an optimized binary algorithm. By using the binary search strategy in optimization theory, combined with information entropy and coefficient of variation theory, the threshold value is dynamically adjusted. The initial threshold value is set to 80% of the resource capacity as a reference point. An adaptive adjustment mechanism is introduced to dynamically correct the threshold value according to the real-time changes of load balancing degree. The coefficient of variation is used to evaluate the dispersion degree of resource load, and the entropy weight method is used to calculate the uncertainty of resource allocation. In the binary search process, the load balancing index and resource utilization rate are calculated in each iteration to gradually converge to the optimal threshold value. The convergence threshold is set to 0.05, i.e. the threshold value changes less than 5% between adjacent iterations to stop searching. For complex multi-dimensional resource scenarios, a multi-dimensional binary search strategy is used to optimize the threshold value of different resource dimensions in parallel. Random disturbance terms and bootstrap sampling techniques are introduced to enhance the robustness and global search ability of the algorithm.

[0051] The specific implementation of step S80 is to perform microservice isolation operations. A real-time monitoring and migration system based on event-driven is constructed, and complex event processing (CEP) technology is used to continuously monitor the resource unit load. When the load exceeds the preset threshold, a multi-stage migration decision-making process is triggered. First, the Monte Carlo simulation method is used to evaluate the migration cost, considering the state migration data volume, network channel occupancy rate, migration time and service impact degree. The Markov decision process (MDP) model is introduced to model the migration problem as a sequential decision optimization problem. The Q-learning algorithm in reinforcement learning is used to learn the optimal migration strategy. The migration execution path selection is based on the microservice correlation matrix and the minimum cost path algorithm in graph theory. A smooth migration mechanism based on container orchestration technology is designed to ensure zero downtime migration of services. By introducing dynamic threshold and fault tolerance mechanism, the migration cost and service performance are effectively balanced, realizing fine-grained resource isolation and dynamic scheduling of microservices.

[0052] The second aspect of the present application provides a computer readable storage medium, wherein program instructions are stored in the computer readable storage medium, and the program instructions are used to execute the productivity platform microservice isolation running method when running in a computer.

[0053] The third aspect of the present application provides a productivity platform microservice isolation running system, comprising the computer readable storage medium described above, the system is any one of a computer, a server, and a single-chip microcomputer, the computer readable storage medium is arranged in the system, and the system is provided with a microprocessor for executing the program instructions stored in the computer readable storage medium.

[0054] Specifically, the principle of the present application is that the technical solution of the present application is based on the dynamic equilibrium theory of complex systems, and a multi-dimensional and multi-level resource analysis and scheduling model is constructed. The core principles include multi-scale load feature decomposition, dynamic resource matrix construction, resource competition index quantification, and deep analysis of service correlation, and other advanced technologies.

[0055] The scheme accurately describes the resource consumption characteristics of the microservice by introducing mathematical tools such as information entropy and coefficient of variation. Advanced algorithms such as tensor decomposition and wavelet transform are used to realize multi-dimensional and high-precision modeling of the resource behavior of complex distributed systems. Through the construction of a dynamic and adaptive resource allocation equation set, the precise and real-time regulation of microservice resources is realized.

[0056] The core logic of the technical solution is to transform the microservice resource allocation problem into a quantifiable and dynamically optimized system engineering problem. By establishing a resource management model based on big data and artificial intelligence, the technical bottleneck of traditional static resource allocation is broken through, and the intelligentization, precision and dynamic adaptation of resource allocation are realized.

[0057] A specific embodiment 1 of the present application is provided below, and the specific implementation of each step in the embodiment 1 is described in detail as follows.

[0058] The specific implementation of step S10 is to realize comprehensive acquisition of microservice running index data through a multi-dimensional and high-precision index collection system. The data collection process follows the principles of hierarchical collection and intelligent filtering. For processor occupancy and memory occupancy, an accurate sampling method based on hardware performance counters is used, specifically through real-time capture of processor clock cycles and memory allocation through the underlying interface of the operating system. The monitoring of network throughput and disk read / write rate uses tracking technology of the underlying network protocol stack and storage subsystem, which can accurately obtain the instantaneous rate of data transmission and storage. The statistics of service response delay and service call frequency rely on a distributed tracking system, and through the implantation of a lightweight tracking component in the service call link, the timestamps, input parameters, output parameters and other key information of each service call are recorded.

[0059] The accurate measurement of data exchange volume requires the design of a special statistical model. The data exchange volume statistical model is defined as follows:

[0060]

[0061] where V ij represents the data exchange volume between service i and service j, S ik represents the data size sent by service i, R jk represents the data size received by service j, and n is the number of exchanges in the statistical period.

[0062] The environmental parameter data is collected by a multi-sensor array deployed in the server room, including temperature, humidity, air pressure, and other physical indicators. The deployment of the sensor array follows the principle of spatial uniform distribution to ensure the representativeness and accuracy of the data. For the collected raw index data, standardization and normalization processing is needed to eliminate the influence of different index dimensions.

[0063] The calculation of microservice priority indicators uses a multi-level analysis method. The priority calculation model is constructed as follows:

[0064] P i =ω1·ρ i +v2·γ i +ω3·η i ;

[0065] where P i is the priority indicator of microservice i, ρ i is the service importance coefficient, γ i is the service load contribution coefficient, η i is the service environmental adaptability coefficient, ω1, ω2, and ω3 are the corresponding weight coefficients, and satisfy ω1+ω2+ω3=1. The weight coefficients are determined by the Analytic Hierarchy Process (AHP), and the reference value range is 0.3 to 0.4.

[0066] The specific implementation of step S20 is to use time series decomposition theory to perform complex decomposition operations on microservice running indicator data. Based on the given decomposition formula:

[0067] L total =L base +L dynamic +∈;

[0068] where the error term ∈ is in the range of 0.01 to 0.05. The decomposition process uses multi-scale wavelet transform theory to decompose the total load signal into sub-signals of different scales and frequencies. The basic load L base is extracted using an improved Holt's exponential smoothing method, which can effectively remove short-term fluctuations and capture long-term trends. The basic load extraction model is defined as follows:

[0069] L base(t) = a · L total(t) + (1 - a) · L base(t-1) ;

[0070] where a is a smoothing coefficient, taking a value in the range 0.1 to 0.3, and t represents a discrete time point. The dynamic load L dynamic is extracted using a multi-scale decomposition technique based on wavelet transform. A suitable wavelet basis function (such as Haar wavelet) is selected, and short-term fluctuation components are obtained through multi-level decomposition.

[0071] The dynamic load extraction model is defined as:

[0072] L dynamic(t) = L total(t) - L base(t) ;

[0073] To improve the robustness of decomposition, an outlier detection mechanism based on kernel density estimation is introduced. The original load data is preprocessed, and extreme data points are identified and removed using the Gaussian kernel density estimation method.

[0074] The specific implementation of step S30 is to construct a resource matrix with multi-dimensional semantics. The benchmark resource matrix M base is constructed using tensor decomposition theory, which maps multi-dimensional heterogeneous data to a low-dimensional semantic space. Based on the given matrix expression:

[0075]

[0076] where r ij represents the basic load value of the i-th microservice on the j-th resource, m is the number of microservices, and n is the number of resource types. Principal component analysis (PCA) is used for dimensionality reduction, and the main components with an explanation variance contribution rate of more than 95% are retained. The PCA dimensionality reduction model is defined as:

[0077]

[0078] Subject to:

[0079] where λ j is the eigenvalue, and k is the number of retained principal components.

[0080] The construction of the dynamic resource matrix M dynamic introduces a dynamic tensor decomposition method. Based on the given matrix expression:

[0081]

[0082] The multi-dimensional data is decomposed into a low-rank core tensor and factor matrices by using CANDECOMP / PARAFAC (CP) decomposition algorithm. The CP decomposition model is defined as:

[0083]

[0084] where X is the input tensor, λ r is the weight coefficient, a r , b r , c r are factor matrices, R is the rank of decomposition, denotes the Kronecker product.

[0085] The specific implementation of step S40 is to calculate the competition index through complex statistical and information theory methods. Based on the given competition index calculation formula:

[0086]

[0087]

[0088] The calculation of the instantaneous competition index C instant introduces the coefficient of variation and Shannon entropy theory. The coefficient of variation calculation model is defined as:

[0089]

[0090] The Shannon entropy weight calculation model is:

[0091]

[0092] where is the normalized probability.

[0093] The calculation of the persistent competition index C persist uses the kernel density estimation method. The Gaussian kernel density estimation model is defined as:

[0094]

[0095] where K is the Gaussian kernel function, h is the bandwidth parameter, and is usually 1.06 times the empirical standard deviation. The adjustment coefficients α, β, γ, δ have a value range of 0 to 1, and the optimal value can be determined by cross-validation and grid search methods.

[0096] The specific implementation of step S50 is to construct a microservice call graph, and the calling relationship between services is abstracted into a directed weighted graph using complex network analysis theory. The call graph construction process is divided into multiple key sub-steps. First, data collection is performed, and detailed metadata of microservice calls are collected through a distributed tracing system (such as OpenTelemetry). The service correlation degree calculation model is defined as follows:

[0097]

[0098] where A ij represents the correlation degree between microservices i and j, f ij represents the service call frequency, v ij is the data exchange volume, t ij represents the service response delay, s ij is the correlation strength index, ξ is the noise coefficient. The correlation strength s ij The calculation introduces the complex network centrality theory:

[0099]

[0100] When constructing the correlation matrix, the compression storage technology is adopted for efficient storage of sparse matrices. The graph embedding technology is introduced to map the high-dimensional service correlation relationship to the low-dimensional semantic space. Specifically, the distributed representation of service nodes is learned through the random walk algorithm (such as DeepWalk) to capture the topological structure and semantic relationship between services.

[0101] The specific implementation of step S60 is to construct a resource allocation equation set, and to convert the resource allocation problem into a Pareto optimal problem using multi-objective optimization theory. The utility calculation equation combines the theories of game theory and information theory:

[0102]

[0103] where U ij is the utility index of microservice i in resource unit j, P i is the microservice priority, R ij is the resource request vector, C j is the resource unit capacity, λ is the time decay coefficient, κ is the resource change rate weight, and μ is the random disturbance term.

[0104] The equation solving adopts an improved particle swarm optimization algorithm combined with simulated annealing technology for global optimal solution search. The Lagrange dual theory is introduced to convert the complex multi-objective optimization problem into an equivalent dual problem. The specific solving process includes: initial population random generation, fitness evaluation, particle velocity and position update, local and global optimal solution search, etc.

[0105] The specific implementation of step S70 is to determine the resource usage threshold value using the optimized bisection algorithm. The algorithm design is based on optimization theory and information entropy theory. The initial threshold value is set to 80% of the resource capacity, and an adaptive adjustment mechanism is introduced:

[0106] T j = T0·(1±δ) k ;

[0107] wherein, T j is the adjusted threshold value, T0 is the initial threshold value, δ is the adjustment step, and k is the iteration number. The coefficient of variation and entropy weight theory are introduced to evaluate the balance of resource load allocation. Specifically, by calculating the standard deviation and information entropy of the load, the threshold value is dynamically adjusted.

[0108] The convergence criterion is that the threshold value changes less than 5% between two adjacent iterations. The algorithm also introduces random disturbance and bootstrap sampling techniques to enhance global search capability. For multi-dimensional resource scenarios, a parallel multi-dimensional binary search strategy is used to optimize the threshold values of different resource dimensions simultaneously.

[0109] The specific implementation of step S80 is to perform microservice isolation operations. A real-time monitoring and migration system based on complex event processing (CEP) technology is constructed. A migration cost calculation model is defined:

[0110]

[0111] wherein, M cost is the migration cost index, D is the state migration data volume, W is the network channel occupancy rate, T is the migration time, I is the service impact degree, θ1, θ2, θ3 are weight coefficients, and η is the risk factor.

[0112] The migration decision is comprehensively evaluated by Markov decision process (MDP) and Monte Carlo simulation method. The Q-learning algorithm in reinforcement learning is used to learn the optimal migration strategy. The migration execution path selection is based on the microservice correlation matrix and the minimum cost path algorithm. A smooth migration mechanism based on container orchestration technology is designed to ensure zero downtime migration of services. Through dynamic threshold and fault tolerance mechanism, the migration cost and service performance are balanced to realize fine-grained resource isolation and dynamic scheduling of microservices.

[0113] In order to better understand and implement the present application, the following provides an embodiment 2 of a specific application scenario of the present application: An enterprise is undergoing digital transformation using a productivity platform architecture and has deployed a large number of microservice-based application systems. As the business scale expands rapidly, the number of microservices deployed in the productivity platform of the enterprise has exceeded 1000, covering multiple fields such as financial management, human resources, and supply chain collaboration. To ensure the stable operation of key businesses, the enterprise urgently needs to realize automatic isolation management of these microservices.

[0114] Based on this, the enterprise decides to use the productivity platform microservice isolation running method proposed by the present application. The following takes the implementation of the enterprise as an example to introduce the specific application process of the method in detail.

[0115] First, the IT operation team of the enterprise collects the running index data of all microservices. Through the monitoring system, the IT team collects the real-time load of each microservice in the dimensions of CPU, memory, network, disk, etc. Taking a certain financial management microservice as an example, the index data of the microservice in the past 7 days is shown in Table 1.

[0116] Table 1 Running index data of a certain financial management microservice in the past 7 days

[0117]

[0118]

[0119] Figure 2 The performance indicators of the financial management microservice in the dimensions of CPU occupancy rate, memory occupancy rate, network throughput, disk read / write rate, response delay, and call frequency are shown. The relative level of each indicator is intuitively presented in the form of a radar chart. According to these index data, the operation team calculates the comprehensive priority indicator P of the microservice i = 0.8. This indicator comprehensively considers the performance, importance, and environment of the microservice, reflecting the position of the microservice in the entire productivity platform.

[0120] Then, the operation team calculates the running index data of all microservices. First, the total load L total of each microservice is divided into basic load L base and dynamic load L dynamic . Taking the above financial management microservice as an example, after analyzing the index data of the microservice in the past 7 days, L base = 49%, L dynamic = 3%, and ∈ = 0.02 are obtained. This indicates that 49% of the resource demand of the microservice is stable basic demand, 3% of the demand is dynamic fluctuation, and the overall load fluctuation is within 2%. Figure 3 The basic load, dynamic load, and priority indicator of different microservices are shown in combination with bar charts and line charts.

[0121] Based on the above load decomposition results, the operation team constructs the benchmark resource matrix M base and the dynamic resource matrix M dynamic . Taking the CPU resource as an example, the data of a certain row of M base is shown in Table 2:

[0122] Table 2 Part of the data of M base

[0123] Microservices CPU base load Financial management 49％ HR management 42％ Supply chain 55％ Project management 38％ … …

[0124] And M dnamic ​The data in the row is shown in Table 3 below:

[0125] Table 3 M dynamic Part of the data

[0126]

[0127]

[0128] By constructing these two matrices, the operation and maintenance team fully describes the load characteristics of each microservice in different resource dimensions.

[0129] Next, the operation and maintenance team calculates the instantaneous competition index C instant and the persistent competition index C persist of each microservice. Taking the CPU resource as an example, assuming that the resource weight coefficient w i = 0.4, the adjustment coefficients α = 0.6, β = 0.4, γ = 0.5, δ = 0.5, then:

[0130] The CPU instantaneous competition index C instant = 1.45;

[0131] The CPU persistent competition index C persist = 3.29;

[0132] Similarly, the operation and maintenance team calculates the competition index in other resource dimensions. Figure 4 The relationship between the instantaneous competition index of each microservice and the correlation degree is shown by a scatter plot. The figure includes specific values such as C_{instant} = 1.45 calculated in the example, as well as results such as A_{ij} = 6.77 obtained from the correlation degree calculation equation. These indicators comprehensively reflect the degree of competition for resources by each microservice in the short and long term, providing a basis for subsequent resource allocation decisions.

[0133] Based on the above microservice resource characteristic analysis results, the operation and maintenance team constructs a resource allocation equation set.

[0134] First, the utility calculation equation U ij is used to evaluate the utility index of microservices in different resource units:

[0135]

[0136] where R ij is the number of requests of the i-th microservice on the j-th resource, C j is the capacity of the j-th resource unit, and t is the running time. This equation takes into account factors such as microservice priority, timeliness and dynamics of resource usage, and can more accurately evaluate the resource needs of microservices. Figure 5 The utility calculation equation Uij The relationship between the runtime, resource request ratio, and utility index.

[0137] Correlation calculation equation A ij Used to evaluate the degree of correlation between microservice i and microservice j: A ij = 6.77;

[0138] Where 1200 is the microservice call frequency, 500 is the data exchange volume, 150 is the service response delay, 0.8 is the correlation strength index, and 0.1 is the noise coefficient. This equation balances the influence of multiple performance indicators and more comprehensively describes the correlation between microservices.

[0139] Load balancing equation B j Used to ensure the balance of resource allocation:

[0140]

[0141] Where L ij is the load index of the ith microservice on the jth resource unit, H j is the load change record of the jth resource unit, and σ is the standard deviation function. This equation considers resource usage intensity, load fluctuation, and load change trend, effectively maintaining the balance of resource allocation.

[0142] Finally, migration cost equation M cost Used to evaluate the overhead of microservice migration: M cost = 59.78;

[0143] Where 500 is the state migration data volume, 90 is the network channel occupancy rate, 180 is the migration time, 2 is the service impact degree, and 0.05 is the risk factor. This equation comprehensively evaluates the various overheads in the migration process, providing a basis for migration decisions. Figure 6 The migration cost changes of different priority microservices under different resource utilization rates are shown, and the resource usage threshold T

[0144] By solving the above resource allocation equation set, the operation and maintenance team obtains a resource allocation scheme that meets the correlation between microservices and load balancing. Next, the IT team uses the optimized binary algorithm to determine a reasonable resource usage threshold T j . Taking CPU resources as an example, the initial setting is T CPU = 80%, and then through iterative optimization, the final T CPU = 75% is determined.

[0145] With the resource allocation scheme and reasonable usage threshold, the operation and maintenance team begins to monitor the load of each resource unit in real time. For those resource units whose load exceeds the threshold, the team calculates the migration cost M cost of the microservices and selects appropriate target resource units for migration according to the microservice correlation matrix A.

[0146] For example, in a certain monitoring, it is found that the load of the CPU resource unit exceeds 75%, and the calculation M cost = 59.78 meets the expected requirements. Therefore, the team decides to migrate the financial management microservice from this resource unit to another resource unit with lower load. Through this dynamic isolation operation, the operation and maintenance team can timely alleviate resource bottlenecks and ensure the stable operation of critical business services.

[0147] In the following operation process, the enterprise's IT team continuously monitors the resource usage of the productivity platform, and timely adjusts the resource allocation and usage threshold. At the same time, the IT team also observes the load change trend of each microservice, and timely optimizes the parameters of the resource allocation equation set to improve the accuracy of resource management.

[0148] For example, with the seasonal fluctuations of business, the dynamic load of some microservices has increased significantly. According to historical data analysis, the operation and maintenance team finds that the time decay coefficient λ of such microservices should be appropriately reduced to increase the weight of their short-term resource demand. At the same time, the IT team also appropriately increases the weight coefficient φ of load change acceleration in the load balancing equation to enhance the early warning ability for sharp fluctuations.

[0149] Through continuous optimization, the enterprise's productivity platform microservice isolation operation scheme is constantly improved, and the resource utilization efficiency and service stability are continuously improved. At present, the enterprise has basically realized the automatic isolation management of microservices, and the average availability of critical business services has reached 99.9%, which is much higher than before. The IT team has also received high recognition from the top management, and has made important contributions to the digital transformation of the enterprise.

[0150] It should be noted that the variables involved in the present application are explained in detail as shown in Table 4.

[0151] Table 4 Variable Explanation Table

[0152]

[0153]

[0154] The above description 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 can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for isolating and running microservices in a productivity middleware platform, characterized in that, The microservice running index data is collected, the microservice running index data is decomposed to obtain a basic load value and a dynamic load value, a benchmark resource matrix is established based on the basic load value, a dynamic resource matrix is established based on the dynamic load value, resource unit capacity data is generated according to the benchmark resource matrix, an instantaneous competition index is calculated based on the dynamic resource matrix, a persistent competition index is calculated based on the benchmark resource matrix, a resource request vector is generated in combination with the instantaneous competition index and the persistent competition index, a microservice call graph is constructed, a microservice correlation matrix is established, a microservice grouping is converted into a multi-dimensional constraint knapsack problem, and a resource allocation equation group including a utility calculation equation, a correlation degree calculation equation, a load balancing equation and a migration cost equation is used to solve a resource allocation scheme to obtain a resource balance constraint index, a resource usage threshold value is determined based on the resource balance constraint index, and resource units are divided, the resource unit load index is monitored, and a microservice isolation operation is performed. 2.The productivity platform micro-service isolation running method according to claim 1, characterized in that, The microservice running index data includes processor occupancy, memory occupancy, network throughput, disk read / write rate, service response delay, service call frequency, data exchange volume, and environmental parameter data, and the microservice priority index is generated by the microservice running index data. 3.The productivity platform micro-service isolation running method according to claim 2, characterized in that, The correlation strength index between microservices is analyzed by path optimization operation, and the microservice correlation matrix is established. 4.The productivity platform micro-service isolation running method according to claim 3, characterized in that, The resource allocation equation group includes the utility calculation equation, the correlation degree calculation equation, the load balancing equation, and the migration cost equation. 5.The productivity platform micro-service isolation running method according to claim 4, characterized in that, The inputs of the utility calculation equation include the resource request vector, the resource unit capacity data, and the microservice priority index, and the output is a microservice utility allocation matrix; the inputs of the correlation degree calculation equation include the service call frequency, the data exchange volume, the correlation strength index, and the service response delay, and the output is the microservice correlation matrix. 6.The productivity platform micro-service isolation running method according to claim 5, characterized in that, The inputs of the load balancing equation include the resource unit load index, the load change record, the microservice priority index, and the resource usage threshold value, and the output is the resource balance constraint index; the inputs of the migration cost equation include the state migration data volume, the network channel occupancy rate, the migration operation time, and the service influence degree, and the output is the migration cost index.

7. The method of claim 6, wherein the productivity platform microservice isolation running method is characterized by, When the resource unit load index exceeds the resource usage threshold value, a migration execution path is determined based on the state migration data volume, the network channel occupancy rate, the migration operation time, and the service influence degree, and the microservice isolation operation is completed.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores program instructions, which when running in the computer, are used to execute the productivity platform microservice isolation running method of any one of claims 1-7.

9. A productivity platform microservice isolated running system, characterized in that, The computer readable storage medium of claim 8, the system is any one of a computer, a server, and a single chip microcomputer, the computer readable storage medium is arranged in the system, and the system is provided with a microprocessor for executing the program instructions stored in the computer readable storage medium.

Images