SaaS application component upgrading method and system

Abstract

The invention belongs to the technical field of software upgrading, and discloses a software as a service (SaaS) application component upgrading method and system. The method comprises the following steps: in an SaaS platform, constructing a data analysis model, a resource scheduling model and an upgrading strategy generation model; collecting real-time monitoring data, and performing standardization processing; performing data analysis by using a data analysis model according to standard real-time monitoring data; performing resource scheduling by using a resource scheduling model according to a real-time data analysis result; generating an upgrading strategy by using an upgrading strategy generation model according to the real-time data analysis result and the real-time resource scheduling scheme; and performing offline upgrading on the corresponding SaaS application component according to the real-time upgrading strategy. According to the method, the problems of high cost, low practicability, lack of flexibility and adaptivity, low resource utilization rate and poor strategy dynamism in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of software upgrade, and in particular relates to a SaaS application component upgrade method and system. Background Art

[0002] Software as a Service (SaaS) application components refer to the different parts or modules that make up the application in the SaaS platform, which work together to provide the complete services of the SaaS platform. According to the architecture and requirements of the SaaS platform, the deployment types of SaaS application components include independent components, integrated components, and hybrid components. SaaS application component upgrade is an important part of software maintenance and development, which involves updating existing software components to introduce new features, improve performance, fix vulnerabilities, or enhance user experience.

[0003] The existing SaaS application component upgrade method has the following defects:

[0004] 1) High cost and low practicality: Traditional SaaS application component upgrades often rely on manual operations, which is not only time-consuming and labor-intensive, but also error-prone. Due to manual intervention, omissions or configuration errors may occur during the upgrade process, resulting in service interruptions or performance degradation.

[0005] 2) Lack of flexibility and adaptability: Existing upgrade methods cannot be personalized according to the characteristics and needs of different application components. This inflexible upgrade process may not be suitable for all components, resulting in resource waste or poor upgrade results.

[0006] 3) Low resource utilization: Existing upgrade methods do not fully consider resource usage, resulting in uneven resource allocation and low resource utilization during the upgrade process, which increases operating costs.

[0007] 4) Poor policy dynamics: Existing upgrade strategies are often pre-defined and cannot be dynamically adjusted according to real-time system status and external environment changes, which may lead to a mismatch between upgrade strategies and actual needs. Summary of the invention

[0008] In order to solve the problems of high cost and low practicality, lack of flexibility and adaptability, low resource utilization and poor policy dynamics in the prior art, the present invention aims to provide a SaaS application component upgrade method and system.

[0009] The technical solution adopted by the present invention is:

[0010] A SaaS application component upgrade method comprises the following steps:

[0011] In the SaaS platform, data analysis models, resource scheduling models, and upgrade strategy generation models are built based on historical monitoring data of different SaaS application components;

[0012] Collect the real-time monitoring data of SaaS application components, and standardize the real-time monitoring data to obtain standard real-time monitoring data of SaaS application components;

[0013] Monitor data in real time according to standards, use data analysis models to perform data analysis, and obtain real-time data analysis results of SaaS application components;

[0014] According to the real-time data analysis results, use the resource scheduling model to perform resource scheduling and obtain the real-time resource scheduling solution for SaaS application components;

[0015] According to the real-time data analysis results and the real-time resource scheduling plan, the upgrade strategy generation model is used to generate the upgrade strategy and obtain the real-time upgrade strategy of the SaaS application component;

[0016] According to the real-time upgrade strategy, the corresponding SaaS application components are upgraded offline to obtain the upgraded SaaS application components, and then return to the monitoring data collection step.

[0017] Furthermore, the historical / real-time monitoring data includes historical / real-time online operation data, historical / real-time offline log data, and historical / real-time basic attribute data of the SaaS application components.

[0018] Furthermore, in the SaaS platform, based on the historical monitoring data of different SaaS application components, a data analysis model, a resource scheduling model, and an upgrade strategy generation model are constructed, including the following steps:

[0019] In the SaaS platform, a unified data model is built based on the common data format supported by the SaaS platform;

[0020] Collect some historical monitoring data of different SaaS application components, and pre-process some heterogeneous historical monitoring data to obtain some pre-processed historical monitoring data;

[0021] According to the unified data model, set corresponding data mapping and format conversion strategies for different SaaS application components;

[0022] According to the data mapping and format conversion strategy, the corresponding pre-processed historical monitoring data is mapped and format converted to obtain a number of isomorphic standard historical monitoring data;

[0023] Based on several standard historical monitoring data, a deep learning algorithm is used to build a data analysis model and generate several historical data analysis results;

[0024] Based on the analysis results of several historical data, a swarm intelligence optimization algorithm is used to build a resource scheduling model and generate several historical resource scheduling plans;

[0025] According to several historical data analysis results and corresponding historical resource scheduling plans, a reinforcement learning algorithm is used to build an upgrade strategy generation model and generate several historical upgrade strategy generation experiences.

[0026] Furthermore, the data analysis model is constructed based on the BiLSTM-Attention-RF algorithm, and the data analysis model includes an input layer connected in sequence, an attention weight layer constructed based on the Attention mechanism, a data feature extraction layer constructed based on the BiLSTM algorithm, a data analysis prediction layer constructed based on the RF algorithm, and an output layer;

[0027] The resource scheduling model is constructed based on the ICPO algorithm, and the resource scheduling model includes an optimization target generation module, an optimal solution search module, and a solution generation module which are connected in sequence;

[0028] The upgrade strategy generation model is built based on the DQN algorithm, and the upgrade strategy generation model includes an agent, a deep Q network, and an experience replay pool.

[0029] Furthermore, real-time monitoring data of SaaS application components are collected and standardized to obtain standard real-time monitoring data of SaaS application components, including the following steps:

[0030] Collect real-time monitoring data of SaaS application components in the SaaS platform, and pre-process the real-time monitoring data to obtain pre-processed real-time monitoring data;

[0031] According to the data mapping and format conversion strategy of the SaaS application component, the pre-processed real-time monitoring data is subjected to data mapping and format conversion to obtain the standard real-time monitoring data of the SaaS application component.

[0032] Furthermore, according to the standard real-time monitoring data, the data analysis model is used to perform data analysis to obtain the real-time data analysis results of the SaaS application components, including the following steps:

[0033] Perform sequence conversion on the standard real-time monitoring data to obtain an input data sequence, and input the input data sequence into the input layer of the data analysis model;

[0034] According to the attention weight value preset in the attention weight layer, the input data sequence is weighted to obtain a weighted input data sequence, and the weighted input data sequence is input into the data feature extraction layer;

[0035] Use the data feature extraction layer to extract the data feature sequence of the weighted input data sequence, convert the data feature sequence into a data feature vector, and input the data feature vector into the data analysis and prediction layer;

[0036] Using the data analysis and prediction layer, extracting several key data features of the data feature vector, and performing data analysis and prediction based on the several key data features to obtain the data analysis and prediction code;

[0037] Use the output layer to decode the data analysis prediction code to obtain real-time data analysis results of the SaaS application component.

[0038] Furthermore, according to the real-time data analysis results, the resource scheduling model is used to perform resource scheduling to obtain a real-time resource scheduling solution for the SaaS application component, including the following steps:

[0039] Use the optimization target generation module of the resource scheduling model to parse the real-time data analysis results, obtain the real-time optimization target and real-time search space, and generate the real-time objective function based on the real-time optimization target;

[0040] Use the optimal solution search module to iteratively search in the real-time search space according to the real-time objective function to obtain the position encoding vector corresponding to the optimal solution;

[0041] The solution generation module is used to decode the position encoding vector corresponding to the optimal solution to obtain the real-time resource scheduling solution for the SaaS application component.

[0042] Furthermore, according to the real-time data analysis results and the real-time resource scheduling scheme, an upgrade strategy generation model is used to generate an upgrade strategy to obtain a real-time upgrade strategy for the SaaS application component, including the following steps:

[0043] According to the real-time data analysis results and real-time resource scheduling plan, search the experience playback pool of the upgrade strategy generation model to extract several matching historical upgrade strategy generation experiences;

[0044] According to several matching historical upgrade strategy generation experiences, the action space of the upgrade strategy generation model is updated to obtain an updated action space of several possible actions;

[0045] According to the real-time data analysis results and the real-time resource scheduling scheme, the state space of the upgrade strategy generation model is updated to obtain an updated state space including a plurality of states;

[0046] Input the updated state space into the deep Q network, and use the agent to control the deep Q network to output the Q value sequence of possible action sequences in the updated action space;

[0047] According to the preset reward function, the Q value sequence is iteratively updated to obtain an updated Q value sequence until the number of iterations reaches the threshold;

[0048] According to the greedy strategy, the possible action with the highest updated Q value in the updated Q value sequence of each iteration is used as the execution action of the corresponding state;

[0049] The execution actions of all states in the updated state space are integrated to obtain the real-time upgrade strategy of SaaS application components.

[0050] Further, the real-time data analysis results include real-time application component status analysis results, real-time upgrade demand analysis results, and real-time upgrade risk analysis results;

[0051] The real-time application component status analysis results include real-time cost analysis results, real-time performance analysis results, real-time resource utilization analysis results, and real-time stability analysis results;

[0052] The real-time resource scheduling solution includes real-time resource demand and real-time resource scheduling strategy;

[0053] The real-time upgrade strategy includes real-time upgrade decision and real-time resource scheduling decision.

[0054] A SaaS application component upgrade system is used to implement a SaaS application component upgrade method. The system is set on a SaaS platform, and the system includes a model building unit, a data collection unit, a data analysis unit, a resource scheduling unit, an upgrade strategy generation unit and an offline upgrade unit connected in sequence. The SaaS platform is provided with a plurality of SaaS application components, and each SaaS application component is respectively connected to the data collection unit and the offline upgrade unit.

[0055] The beneficial effects of the present invention are:

[0056] The present invention discloses a SaaS application component upgrade method and system. By constructing a data analysis model, a resource scheduling model and an upgrade strategy generation model, the intelligent upgrade process is realized, manual intervention is reduced, cost investment is reduced, practicality is enhanced, and upgrade efficiency and accuracy are improved; the data analysis model has the ability to monitor the status of application components in real time and perform real-time analysis, so that problems can be discovered and solved in a timely manner, and upgrade risks are reduced; the resource scheduling model monitors resource usage, intelligently schedules resources, improves resource utilization, reduces operating costs, and can optimize resource allocation during the upgrade process, minimize business interruptions, and ensure user experience and business continuity; the upgrade strategy generation model generates upgrade strategies according to data analysis results and resource scheduling plans, so that the upgrade process can adapt to the current status and needs of application components, improves the flexibility and adaptability of the strategy, realizes dynamic generation of upgrade strategies, and can be adjusted according to real-time system status and environmental changes to ensure that the upgrade strategy always matches actual needs.

[0057] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flowchart of the SaaS application component upgrade method in the present invention.

[0059] Figure 2 It is a structural block diagram of the SaaS application component upgrade system in the present invention. DETAILED DESCRIPTION

[0060] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments.

[0061] Embodiment 1:

[0062] like Figure 1 As shown, this embodiment provides a SaaS application component upgrade method, including the following steps:

[0063] S1: In the SaaS platform, based on the historical monitoring data of different SaaS application components, a data analysis model, a resource scheduling model, and an upgrade strategy generation model are constructed, including the following steps:

[0064] S1-1: In the SaaS platform, a unified data model is constructed according to the common data format supported by the SaaS platform, including the following steps:

[0065] S1-1-1: In the SaaS platform, several core data elements are set according to the common data formats supported by the SaaS platform;

[0066] S1-1-2: Based on several core data elements, set the data structure of the unified data model, such as tables, fields, relations, etc.;

[0067] S1-1-3: Based on several core data elements, set the data relationships between the core data elements of the unified data model, including one-to-one, one-to-many, many-to-many and other relationships;

[0068] S1-1-4: According to several core data elements, the data structure of the unified data model and the data relationship between several core data elements, set the data constraints of each core data element, such as data type, value range, default value, etc.;

[0069] S1-1-5: Construct a corresponding unified data model according to a number of core data elements, the data structure of the unified data model, the data relationship between the core data elements, and the data constraints of each core data element;

[0070] S1-2: collect some historical monitoring data of different SaaS application components, and pre-process some heterogeneous historical monitoring data to obtain some pre-processed historical monitoring data;

[0071] Preprocessing includes data cleaning, duplicate elimination, normalization, and data dimensionality reduction in order to improve data quality and scale the data to a uniform scale to improve the speed and effect of model training. Principal Component Analysis (PCA) is used to reduce data dimensionality. The original data is transformed into a set of linearly independent representations in each dimension through linear transformation to remove noise and redundancy in the data, making the data easier to understand and process.

[0072] Historical monitoring data includes historical online operation data, historical offline log data, and historical basic attribute data of SaaS application components;

[0073] Online operation data includes the operation traffic data, resource usage, operation memory resource usage data, operation network resource usage data, operation cache resource usage, etc. of SaaS application components; offline log data includes the operation log data, security monitoring and protection logs, login logs, etc. of SaaS application components; basic attribute data includes component version information, performance indicator information, compatibility information, etc.

[0074] SaaS application components of SaaS platforms usually include user interface components (dashboards, configuration panels, interactive tools), business logic components (business rule engines, workflow managers, service APIs), data management components (database servers, data models, data synchronization and backup components), infrastructure components (servers / cloud services, load balancers), security components (authentication and authorization components, encryption components, security audit components), performance monitoring and optimization components (performance monitoring tools, log management components), etc.

[0075] S1-3: According to the unified data model, set corresponding data mapping and format conversion strategies for different SaaS application components;

[0076] S1-4: According to the data mapping and format conversion strategy, data mapping and format conversion are performed on the corresponding pre-processed historical monitoring data to obtain a number of isomorphic standard historical monitoring data;

[0077] Data mapping and conversion strategies include mapping strategies and data conversion strategies. Mapping strategies are used to determine how to map heterogeneous data into a unified data model. Data conversion strategies are used to execute preset conversion algorithms or scripts to convert heterogeneous data into data in a unified data model according to mapping strategies. Data mapping and conversion strategies are used to handle data heterogeneity and resolve inconsistencies in different data formats, data types, data semantics, etc.

[0078] S1-5: Based on several standard historical monitoring data, use deep learning algorithms to build a data analysis model and generate several historical data analysis results;

[0079] The data analysis model is built based on the Bidirectional Long Short-Term Memory (BiLSTM)-Attention-Random Forest (RF) algorithm, and includes a sequentially connected input layer, an attention weight layer built based on the Attention mechanism, a data feature extraction layer built based on the BiLSTM algorithm, a data analysis prediction layer built based on the RF algorithm, and an output layer;

[0080] The historical data analysis results include historical application component status analysis results, historical upgrade demand analysis results, and historical upgrade risk analysis results;

[0081] The historical application component status analysis results include historical cost analysis results, historical performance analysis results (such as response time and throughput scores), historical resource utilization analysis results (such as CPU, memory, storage, and network resource utilization), and historical stability analysis results;

[0082] S1-6: Based on the analysis results of several historical data, a swarm intelligence optimization algorithm is used to build a resource scheduling model and generate several historical resource scheduling solutions;

[0083] The resource scheduling model is constructed based on the Improved Crested Porcupine Optimizer (ICPO) algorithm, and the resource scheduling model includes an optimization target generation module, an optimal solution search module, and a solution generation module which are connected in sequence;

[0084] The historical resource scheduling scheme includes historical resource demand conditions and historical resource scheduling strategies;

[0085] S1-7: Based on several historical data analysis results and corresponding historical resource scheduling schemes, a reinforcement learning algorithm is used to build an upgrade strategy generation model and generate several historical upgrade strategy generation experiences, including the following steps:

[0086] S1-7-1: Use the SaaS application component upgrade strategy generation problem as the simulation environment for the DQN algorithm to build an intelligent agent and experience replay pool;

[0087] S1-7-2: Define the action space of the DQN algorithm based on the historical data analysis results and the corresponding historical resource scheduling scheme; the action space includes the state parameters corresponding to the historical data analysis results and the state parameters corresponding to the historical resource scheduling scheme;

[0088] S1-7-3: Define the action space of the DQN algorithm based on the actions that the upgrade strategy needs to output; the action space includes SaaS application component upgrade actions and resource scheduling actions;

[0089] S1-7-4: Based on the possible impact of each action in the action space, define the reward function of the DQN algorithm to evaluate the quality or impact of the action;

[0090] S1-7-5: Construct the input layer, several hidden layers, and output layer of the deep Q network, connect the input layer to the state space, and connect the output layer to the action space;

[0091] S1-7-6: Based on the state space, action space and reward function, according to several historical data analysis results and corresponding historical resource scheduling plans, the deep Q network and intelligent agent are optimized and trained, an upgrade strategy generation model is constructed, and several historical upgrade strategy generation experiences are generated;

[0092] S1-7-7: Store the experience generated by several historical upgrade strategies into the experience replay pool;

[0093] The upgrade strategy generation model is built based on the Deep Q Network (DQN) algorithm, and the upgrade strategy generation model includes an agent, a deep Q network, and an experience replay pool;

[0094] S2: collecting real-time monitoring data of SaaS application components and performing standard processing on the real-time monitoring data to obtain standard real-time monitoring data of SaaS application components, including the following steps:

[0095] S2-1: collect real-time monitoring data of SaaS application components in the SaaS platform, and pre-process the real-time monitoring data to obtain pre-processed real-time monitoring data;

[0096] Real-time monitoring data includes real-time online operation data, real-time offline log data, and real-time basic attribute data of SaaS application components;

[0097] Preprocessing includes normalization and data dimension reduction in sequence;

[0098] S2-2: According to the data mapping and format conversion strategy of the SaaS application component, the pre-processed real-time monitoring data is mapped and format converted to obtain standard real-time monitoring data of the SaaS application component;

[0099] S3: Perform data analysis based on standard real-time monitoring data and use data analysis models to obtain real-time data analysis results of SaaS application components, including the following steps:

[0100] S3-1: Convert the standard real-time monitoring data into a sequence to obtain an input data sequence, and input the input data sequence into the input layer of the data analysis model;

[0101] S3-2: weighting the input data sequence according to the attention weight value preset in the attention weight layer to obtain a weighted input data sequence, and inputting the weighted input data sequence into the data feature extraction layer;

[0102] S3-3: Use the data feature extraction layer to extract the data feature sequence of the weighted input data sequence, convert the data feature sequence into a data feature vector, and input the data feature vector into the data analysis and prediction layer;

[0103] S3-4: Using the data analysis and prediction layer, extracting several key data features of the data feature vector, and performing data analysis and prediction based on the several key data features to obtain the data analysis and prediction code, including the following steps:

[0104] S3-4-1: Use the RF structure trained by the data analysis and prediction layer to extract the feature contribution of several candidate data features in the data feature vector;

[0105] The formula is:

[0106]

[0107] In the formula, is the feature contribution of the j'th candidate data feature; is the feature contribution of the j'th candidate data feature in the i-th decision tree of the random forest; i is the decision tree indicator; j' is the candidate data feature indicator; n is the total number of decisions;

[0108]

[0109]

[0110] In the formula, GI m GI l GI r" is the Gini index of random forest decision tree node m, node l and node r"; p mk" is the proportion of category k" in decision tree node m; K is the total number of categories; m, l, r" are node indicators; k" is the category indicator;

[0111] S3-4-2: normalizing the feature contributions of several candidate data features to obtain corresponding normalized feature contributions;

[0112] The formula is:

[0113]

[0114] Where VIM j' is the contribution of the feature after normalization; J is the total number of candidate data features;

[0115] S3-4-3: Generate feature selection standard values ​​for several candidate data features based on the normalized feature contribution;

[0116] The formula is:

[0117]

[0118] In the formula, CFC j' The feature selection standard value for the j'th candidate data feature; VIM j" is the normalized feature contribution of the j'th candidate data feature; j' is the indicator of the candidate data feature;

[0119] S3-4-4: According to the feature selection standard value, the candidate data features are sorted in descending order, and the first M candidate data features are selected as key data features to obtain M key data features, where M is the total number of key data features;

[0120] S3-4-5: Perform data analysis and prediction based on the M key data features to obtain a data analysis and prediction code;

[0121] S3-5: Use the output layer to decode the data analysis prediction code to obtain the real-time data analysis results of the SaaS application component;

[0122] Real-time data analysis results include real-time application component status analysis results, real-time upgrade demand analysis results, and real-time upgrade risk analysis results;

[0123] The real-time application component status analysis results include real-time cost analysis results, real-time performance analysis results (such as response time and throughput scores), real-time resource utilization analysis results (such as CPU, memory, storage, and network resource utilization), and real-time stability analysis results;

[0124] S4: Based on the real-time data analysis results, use the resource scheduling model to perform resource scheduling and obtain a real-time resource scheduling solution for the SaaS application component, including the following steps:

[0125] S4-1: Use the optimization target generation module of the resource scheduling model to parse the real-time data analysis results, obtain the real-time optimization target and the real-time search space, and generate the real-time objective function according to the real-time optimization target;

[0126] In this embodiment, if the real-time resource utilization analysis result in the real-time data analysis result is low utilization, resource scheduling is required to improve the abnormal situation of the SaaS application component with low utilization;

[0127] Taking maximizing resource utilization as the real-time optimization goal, the formula of the constructed real-time objective function is:

[0128]

[0129] Where F is the real-time objective function; ζ(j) is the utilization rate of the jth resource, including CPU, memory, storage, and network resources; j is the resource indicator; J is the total number of resources;

[0130] S4-2: Using the optimal solution search module, according to the real-time objective function, iteratively search in the real-time search space to obtain the position encoding vector corresponding to the optimal solution, including the following steps:

[0131] S4-2-1: Use the optimal solution search module to encode the resource scheduling solution into the position encoding vector of the ICPO individual in the ICPO algorithm, and use the real-time objective function as the fitness function of the ICPO individual;

[0132] S4-2-2: Set ICPO population parameters and maximum number of iterations of ICPO algorithm;

[0133] S4-2-3: Based on the position coding vector of the ICPO individual and the ICPO population parameters, the Circle chaotic mapping sequence is used for initialization to generate several initial ICPO individuals of the initial ICPO population;

[0134] The formula is:

[0135]

[0136] In the formula, is the initial ICPO individual of the Circle chaos map, i.e. the initial solution; is the randomly generated initial ICPO individual; i' is the ICPO individual indicator;

[0137] S4-2-4: Introduce a cyclic population reduction mechanism to limit the number of individuals in the ICPO population parameters and obtain updated ICPO population parameters for the next iteration;

[0138] The formula is:

[0139]

[0140] In the formula, S t+1 is the number of individuals in the ICPO population parameter of the t+1th iteration; S t is the number of individuals in the ICPO population parameter of the tth iteration; S min is the minimum number of individuals in the ICPO population parameter; a' is the function evaluation parameter; V is the function evaluation cycle parameter; V max is the maximum function evaluation loop parameter; t is the number of iterations indicator;

[0141] S4-2-5: Calculate the initial fitness value of the initial ICPO individuals in the initial ICPO population according to the fitness function;

[0142] S4-2-6: According to the initial fitness value and the updated ICPO population parameters, the first defense strategy, the second defense strategy, the third defense strategy and the fourth defense strategy are used to update the initial ICPO population to obtain an updated ICPO population;

[0143] The formula for the first defense strategy is:

[0144]

[0145] In the formula, For the updated ICPO individuals within the first defense range; is the initial ICPO individual within the first defense range; τ1 is a random number based on normal distribution; τ2 is a random value in the interval [0,1]; It is the optimal solution within the first defense range; is the vector generated between the true optimal solution within the first defense range and the optimal solution randomly selected from the ICPO population; i' is the ICPO individual indicator; t is the iteration indicator;

[0146] The formula for the second defense strategy is:

[0147]

[0148] In the formula, For the updated ICPO individuals within the second defense range; The initial ICPO individual within the second defense range; is the search upper limit vector of the second defense range; τ3 is a random value in the interval [0,1]; are the r1th and r2th initial ICPO individuals respectively; r1 and r2 are two random integers between [1, S]; is the vector generated between the true optimal solution within the second defense range and the optimal solution randomly selected from the ICPO population;

[0149] The formula for the third defense strategy is:

[0150]

[0151] In the formula, For the updated ICPO individuals within the third defense range; The initial ICPO individual within the third defense range; is the search upper limit vector of the third defense range; are the r2th and r3th initial ICPO individuals respectively; r3 is a random integer between [1, S]; The odor diffusion factor defined for the fitness function; λ t It is a defense factor; Control parameters for search direction;

[0152] The formula for the fourth defense strategy is:

[0153]

[0154] In the formula, For the updated ICPO individuals within the fourth defense range; The initial ICPO individual within the fourth defense range; is the optimal solution within the fourth defense range; τ4 and τ5 are both random values ​​in the interval [0,1]; λt It is a defense factor; Control parameters for search direction; is the average force affecting the search direction; a' is the convergence speed factor;

[0155] S4-2-7: Use the dynamic reverse learning algorithm to perform dynamic reverse learning on the updated ICPO population to generate a dynamic reverse ICPO population;

[0156] The formula is:

[0157]

[0158] In the formula, is the ICPO individual in dynamic reverse direction; γ' is the decreasing inertia coefficient; L max , L min are the maximum and minimum values ​​of the vector space respectively; For the updated ICPO individual;

[0159] S4-2-8: According to the fitness function, the fitness values ​​of all ICPO individuals in the updated ICPO population and the dynamically reversed ICPO population are calculated, and the ICPO individual with the maximum fitness value is taken as the optimal individual, and the optimal individual is retained;

[0160] S4-2-9: If the number of iterations of iterative optimization reaches the maximum number of iterations or the fitness value of the optimal individual meets the requirements, the optimal solution corresponding to the optimal individual is output to obtain the position encoding vector corresponding to the optimal solution;

[0161] S4-3: Use the solution generation module to decode the position encoding vector corresponding to the optimal solution to obtain a real-time resource scheduling solution for the SaaS application component;

[0162] The real-time resource scheduling solution includes real-time resource demand and real-time resource scheduling strategy;

[0163] S5: Based on the real-time data analysis results and the real-time resource scheduling plan, an upgrade strategy generation model is used to generate an upgrade strategy to obtain a real-time upgrade strategy for the SaaS application component, including the following steps:

[0164] S5-1: Based on the real-time data analysis results and the real-time resource scheduling plan, search the experience playback pool of the upgrade strategy generation model to extract several matching historical upgrade strategy generation experiences;

[0165] S5-2: Based on some matching historical upgrade strategy generation experiences, the action space of the upgrade strategy generation model is updated to obtain an updated action space A'=[a'1,...,a' j" ,...,a'I ], where a' j" is the updated j-th "action value", j" is the action indicator, and I is the total number of action space dimensions;

[0166] In this embodiment, the SaaS application component upgrade actions in the action space include no upgrade, minor version upgrade, major version upgrade, partial upgrade, phased upgrade, upgrade after resource optimization, rollback to the previous version, minor version upgrade after allocating more CPU resources, major version upgrade after allocating more memory resources, major version upgrade after dynamic resource expansion, etc., and resource scheduling includes resource allocation, resource release, load balancing, resource reservation, dynamic expansion, priority scheduling, etc.;

[0167] S5-3: According to the real-time data analysis results and the real-time resource scheduling plan, the state space of the upgrade strategy generation model is updated to obtain an updated state space S'=[s'1,...,s' i" ,...,s' I' ], where s' i" is the updated i'th state value, i' is the state indicator, and I' is the total number of state space dimensions;

[0168] In this embodiment, the state parameters corresponding to the real-time data analysis results in the action space include real-time cost analysis score, real-time performance analysis score, real-time resource utilization, real-time stability analysis score, real-time upgrade demand, real-time upgrade risk analysis score, etc. The state parameters corresponding to the real-time resource scheduling scheme include real-time resource demand and real-time resource scheduling strategy;

[0169] S5-4: Update the state space S' = [s'1, ..., s' i' ,...,s' I ] Input the deep Q network and use the agent to control the deep Q network and output the updated action space A'=[a'1,...,a' j' ,...,a' I ] is the Q-value sequence of possible action sequences;

[0170] S5-5: Iteratively update the Q value sequence according to the preset reward function to obtain an updated Q value sequence until the number of iterations reaches the number threshold;

[0171] The formula is:

[0172] Q(s' p' ,a' p' )=(1-α")·Q(s p' ,a p' )+α"·(R(s p' ,a p' ,s'p' )+γ·Q max (s p' ,a p' ))

[0173] In the formula, Q(s' p' ,a' p' ) is the updated state value s' p' and the updated action value a' p' The corresponding updated Q value; Q(s p' ,a p' ) is the state value s p' and action value a p' The corresponding predicted Q value; α" is the learning rate; Q max (s p' ,a p' ) is the highest predicted Q value; p' is the comprehensive indicator; γ is the update parameter;

[0174] S5-6: According to the greedy strategy, the possible action with the highest updated Q value in the updated Q value sequence of each iteration is used as the execution action of the corresponding state;

[0175] S5-7: Integrate the execution actions of all states in the updated state space to obtain the real-time upgrade strategy of the SaaS application component;

[0176] The real-time upgrade strategy includes real-time upgrade decision and real-time resource scheduling decision;

[0177] In this embodiment, the real-time upgrade strategy includes allocating more CPU resources and then upgrading the minor version, allocating CPU resources to the current SaaS application component, and releasing resources to other SaaS application components;

[0178] S6: According to the real-time upgrade strategy, the corresponding SaaS application component is upgraded offline to obtain the upgraded SaaS application component, and the process returns to the monitoring data collection step.

[0179] Embodiment 2:

[0180] like Figure 2 As shown, this embodiment provides a SaaS application component upgrade system for implementing a SaaS application component upgrade method. The system is set on a SaaS platform, and the system includes a model building unit, a data collection unit, a data analysis unit, a resource scheduling unit, an upgrade strategy generation unit, and an offline upgrade unit connected in sequence. The SaaS platform is provided with a plurality of SaaS application components, and each SaaS application component is respectively connected to the data collection unit and the offline upgrade unit;

[0181] A model building unit is used to build a data analysis model, a resource scheduling model, and an upgrade strategy generation model in the SaaS platform based on historical monitoring data of different SaaS application components;

[0182] A data collection unit is used to collect real-time monitoring data of SaaS application components and perform standard processing on the real-time monitoring data to obtain standard real-time monitoring data of SaaS application components;

[0183] The data analysis unit is used to monitor data in real time according to standards, use data analysis models, perform data analysis, and obtain real-time data analysis results of SaaS application components;

[0184] The resource scheduling unit is used to perform resource scheduling based on the real-time data analysis results and use the resource scheduling model to obtain a real-time resource scheduling solution for the SaaS application component;

[0185] An upgrade strategy generation unit is used to generate an upgrade strategy using an upgrade strategy generation model according to real-time data analysis results and real-time resource scheduling solutions, thereby obtaining a real-time upgrade strategy for SaaS application components;

[0186] The offline upgrade unit is used to perform offline upgrade on the corresponding SaaS application component according to the real-time upgrade strategy to obtain the upgraded SaaS application component.

[0187] The present invention discloses a SaaS application component upgrade method and system. By constructing a data analysis model, a resource scheduling model and an upgrade strategy generation model, the intelligent upgrade process is realized, manual intervention is reduced, cost investment is reduced, practicality is enhanced, and upgrade efficiency and accuracy are improved; the data analysis model has the ability to monitor the status of application components in real time and perform real-time analysis, so that problems can be discovered and solved in a timely manner, and upgrade risks are reduced; the resource scheduling model monitors resource usage, intelligently schedules resources, improves resource utilization, reduces operating costs, and can optimize resource allocation during the upgrade process, minimize business interruptions, and ensure user experience and business continuity; the upgrade strategy generation model generates upgrade strategies according to data analysis results and resource scheduling plans, so that the upgrade process can adapt to the current status and needs of application components, improves the flexibility and adaptability of the strategy, realizes dynamic generation of upgrade strategies, and can be adjusted according to real-time system status and environmental changes to ensure that the upgrade strategy always matches actual needs.

[0188] The present invention is not limited to the above optional implementations, and anyone can derive other various forms of products under the enlightenment of the present invention. The above specific implementations should not be understood as limiting the scope of protection of the present invention. The scope of protection of the present invention should be based on the definition in the claims, and the description can be used to interpret the claims.

Claims

1. A SaaS application component upgrade method, characterized by: The steps include: In the SaaS platform, data analysis models, resource scheduling models, and upgrade strategy generation models are built based on historical monitoring data of different SaaS application components; Collect the real-time monitoring data of SaaS application components, and standardize the real-time monitoring data to obtain standard real-time monitoring data of SaaS application components; Monitor data in real time according to standards, use data analysis models to perform data analysis, and obtain real-time data analysis results of SaaS application components; According to the real-time data analysis results, use the resource scheduling model to perform resource scheduling and obtain the real-time resource scheduling solution for SaaS application components; According to the real-time data analysis results and the real-time resource scheduling plan, the upgrade strategy generation model is used to generate the upgrade strategy and obtain the real-time upgrade strategy of the SaaS application component; According to the real-time upgrade strategy, the corresponding SaaS application components are upgraded offline to obtain the upgraded SaaS application components, and then return to the monitoring data collection step.

2. A SaaS application component upgrade method according to claim 1, characterized in that: The historical / real-time monitoring data includes historical / real-time online operation data, historical / real-time offline log data and historical / real-time basic attribute data of SaaS application components.

3. A SaaS application component upgrade method according to claim 1, characterized in that: In the SaaS platform, based on the historical monitoring data of different SaaS application components, a data analysis model, a resource scheduling model, and an upgrade strategy generation model are constructed, including the following steps: In the SaaS platform, a unified data model is built based on the common data format supported by the SaaS platform; Collect some historical monitoring data of different SaaS application components, and pre-process some heterogeneous historical monitoring data to obtain some pre-processed historical monitoring data; According to the unified data model, set corresponding data mapping and format conversion strategies for different SaaS application components; According to the data mapping and format conversion strategy, the corresponding pre-processed historical monitoring data is mapped and format converted to obtain a number of isomorphic standard historical monitoring data; Based on several standard historical monitoring data, a deep learning algorithm is used to build a data analysis model and generate several historical data analysis results; Based on the analysis results of several historical data, a swarm intelligence optimization algorithm is used to build a resource scheduling model and generate several historical resource scheduling plans; According to several historical data analysis results and corresponding historical resource scheduling plans, a reinforcement learning algorithm is used to build an upgrade strategy generation model and generate several historical upgrade strategy generation experiences.

4. A SaaS application component upgrade method according to claim 3, characterized in that: The data analysis model is constructed based on the BiLSTM-Attention-RF algorithm, and the data analysis model includes an input layer connected in sequence, an attention weight layer constructed based on the Attention mechanism, a data feature extraction layer constructed based on the BiLSTM algorithm, a data analysis prediction layer constructed based on the RF algorithm, and an output layer; The resource scheduling model is constructed based on the ICPO algorithm, and the resource scheduling model includes an optimization target generation module, an optimal solution search module and a solution generation module which are connected in sequence; The upgrade strategy generation model is constructed based on the DQN algorithm, and the upgrade strategy generation model includes an agent, a deep Q network and an experience replay pool.

5. A SaaS application component upgrade method according to claim 3, characterized in that: Collecting real-time monitoring data of SaaS application components and standardizing the real-time monitoring data to obtain standard real-time monitoring data of SaaS application components includes the following steps: Collect real-time monitoring data of SaaS application components in the SaaS platform, and pre-process the real-time monitoring data to obtain pre-processed real-time monitoring data; According to the data mapping and format conversion strategy of the SaaS application component, the pre-processed real-time monitoring data is subjected to data mapping and format conversion to obtain the standard real-time monitoring data of the SaaS application component.

6. A SaaS application component upgrade method according to claim 4, characterized in that: According to the standard real-time monitoring data, use the data analysis model to perform data analysis to obtain the real-time data analysis results of the SaaS application components, including the following steps: Perform sequence conversion on the standard real-time monitoring data to obtain an input data sequence, and input the input data sequence into the input layer of the data analysis model; According to the attention weight value preset in the attention weight layer, the input data sequence is weighted to obtain a weighted input data sequence, and the weighted input data sequence is input into the data feature extraction layer; Use the data feature extraction layer to extract the data feature sequence of the weighted input data sequence, convert the data feature sequence into a data feature vector, and input the data feature vector into the data analysis and prediction layer; Using the data analysis and prediction layer, extracting several key data features of the data feature vector, and performing data analysis and prediction based on the several key data features to obtain the data analysis and prediction code; Use the output layer to decode the data analysis prediction code to obtain real-time data analysis results of the SaaS application component.

7. A SaaS application component upgrade method according to claim 4, characterized in that: According to the real-time data analysis results, use the resource scheduling model to perform resource scheduling and obtain the real-time resource scheduling solution for SaaS application components, including the following steps: Use the optimization target generation module of the resource scheduling model to parse the real-time data analysis results, obtain the real-time optimization target and real-time search space, and generate the real-time objective function based on the real-time optimization target; Use the optimal solution search module to iteratively search in the real-time search space according to the real-time objective function to obtain the position encoding vector corresponding to the optimal solution; The solution generation module is used to decode the position encoding vector corresponding to the optimal solution to obtain the real-time resource scheduling solution for the SaaS application component.

8. A SaaS application component upgrade method according to claim 4, characterized in that: According to the real-time data analysis results and the real-time resource scheduling plan, the upgrade strategy generation model is used to generate the upgrade strategy to obtain the real-time upgrade strategy of the SaaS application component, including the following steps: According to the real-time data analysis results and real-time resource scheduling plan, search the experience playback pool of the upgrade strategy generation model to extract several matching historical upgrade strategy generation experiences; According to several matching historical upgrade strategy generation experiences, the action space of the upgrade strategy generation model is updated to obtain an updated action space of several possible actions; According to the real-time data analysis results and the real-time resource scheduling scheme, the state space of the upgrade strategy generation model is updated to obtain an updated state space including a plurality of states; Input the updated state space into the deep Q network, and use the agent to control the deep Q network to output the Q value sequence of possible action sequences in the updated action space; According to the preset reward function, the Q value sequence is iteratively updated to obtain an updated Q value sequence until the number of iterations reaches the threshold; According to the greedy strategy, the possible action with the highest updated Q value in the updated Q value sequence of each iteration is used as the execution action of the corresponding state; The execution actions of all states in the updated state space are integrated to obtain the real-time upgrade strategy of SaaS application components.

9. A SaaS application component upgrade method according to claim 8, characterized in that: The real-time data analysis results include real-time application component status analysis results, real-time upgrade demand analysis results and real-time upgrade risk analysis results; The real-time application component status analysis results include real-time cost analysis results, real-time performance analysis results, real-time resource utilization analysis results and real-time stability analysis results; The real-time resource scheduling scheme includes real-time resource demand conditions and real-time resource scheduling strategies; The real-time upgrade strategy includes real-time upgrade decision and real-time resource scheduling decision.

10. A SaaS application component upgrade system, used to implement the SaaS application component upgrade method according to any one of claims 1 to 9, characterized in that: The system is arranged on a SaaS platform, and the system includes a model building unit, a data collection unit, a data analysis unit, a resource scheduling unit, an upgrade strategy generation unit and an offline upgrade unit which are connected in sequence. The SaaS platform is provided with a plurality of SaaS application components, and each of the SaaS application components is respectively connected to the data collection unit and the offline upgrade unit.

Images