CN119621101A - A SaaS application component upgrade method and system - Google Patents
A SaaS application component upgrade method and system Download PDFInfo
- Publication number
- CN119621101A CN119621101A CN202411681986.3A CN202411681986A CN119621101A CN 119621101 A CN119621101 A CN 119621101A CN 202411681986 A CN202411681986 A CN 202411681986A CN 119621101 A CN119621101 A CN 119621101A
- Authority
- CN
- China
- Prior art keywords
- real
- data
- time
- saas application
- upgrade
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/60—Software deployment
- G06F8/65—Updates
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of software upgrading and discloses a method and a system for upgrading a software as a service (SaaS) application component. The method comprises the steps of constructing a data analysis model, a resource scheduling model and an upgrade strategy generation model in a SaaS platform, collecting real-time monitoring data and performing standardization, using the data analysis model to perform data analysis according to the standard real-time monitoring data, using the resource scheduling model to perform resource scheduling according to real-time data analysis results, using the upgrade strategy generation model to perform upgrade strategy generation according to the real-time data analysis results and a real-time resource scheduling scheme, and performing offline upgrade on a corresponding SaaS application component according to the real-time upgrade strategy. The method solves the problems of high cost, low practicability, lack of flexibility and adaptability, low resource utilization rate and poor strategy dynamic property in the prior art.
Description
Technical Field
The invention belongs to the technical field of software upgrading, and particularly relates to a method and a system for upgrading a software as a service (SaaS) application component.
Background
Software as a service (SaaS) application components refer to the different parts or modules that make up the application programs in the SaaS platform, which work together to provide the complete services of the SaaS platform. The deployment types of the SaaS application components include independent components, integrated components, and hybrid components according to the architecture and requirements of the SaaS platform. The upgrading of the SaaS application component is an important link of software maintenance and development, and relates to the updating of the existing software component to introduce new functions, improve performance, repair vulnerabilities or improve user experience.
In the existing SaaS application component upgrading method, the following defects exist:
1) The traditional SaaS application component upgrade is often dependent on manual operation, and the process is time-consuming, labor-consuming and error-prone. Due to manual intervention, omission or configuration errors may occur during the upgrade process, resulting in service interruption or performance degradation.
2) The existing upgrading method cannot be adjusted in a personalized way according to the characteristics and requirements of different application components. Such a inflexible upgrade procedure may not be suitable for all components, resulting in wasted resources or poor upgrade results.
3) The resource utilization rate is low, the existing upgrading method does not fully consider the use condition of the resources, so that the resource is unevenly distributed in the upgrading process, the resource utilization rate is low, and the operation cost is increased.
4) The strategy dynamic property is poor, the existing upgrading strategy is often predefined and cannot be dynamically adjusted according to the real-time system state and external environment change, and the upgrading strategy is possibly not matched with the actual requirement.
Disclosure of Invention
The invention aims to solve the problems of high cost, low practicability, lack of flexibility and adaptability, low resource utilization rate and poor strategy dynamic property in the prior art, and provides a method and a system for upgrading a software as a service (SaaS) application component.
The technical scheme adopted by the invention is as follows:
a SaaS application component upgrading method comprises the following steps:
in the SaaS platform, a data analysis model, a resource scheduling model and an upgrade strategy generation model are built according to historical monitoring data of different SaaS application components;
Collecting real-time monitoring data of the SaaS application assembly, and carrying out standardized processing on the real-time monitoring data to obtain standard real-time monitoring data of the SaaS application assembly;
According to the standard real-time monitoring data, using a data analysis model to perform data analysis to obtain a real-time data analysis result of the SaaS application component;
According to the real-time data analysis result, using a resource scheduling model to perform resource scheduling to obtain a real-time resource scheduling scheme of the SaaS application component;
According to the real-time data analysis result and the real-time resource scheduling scheme, an upgrade strategy generation model is used for generating an upgrade strategy to obtain a real-time upgrade strategy of the SaaS application component;
and carrying out offline upgrading on the corresponding SaaS application assembly according to the real-time upgrading strategy to obtain an upgraded SaaS application assembly, and returning to the monitoring data acquisition step.
Further, the historical/real-time monitoring data includes historical/real-time online running data, historical/real-time offline log data, and historical/real-time basic attribute data of the SaaS application component.
Further, in the SaaS platform, a data analysis model, a resource scheduling model and an upgrade policy generation model are constructed according to the historical monitoring data of different SaaS application components, and the method comprises the following steps:
In the SaaS platform, a unified data model is constructed according to a universal data format supported by the SaaS platform;
Collecting a plurality of historical monitoring data of different SaaS application components, and preprocessing the heterogeneous historical monitoring data to obtain historical monitoring data after intervention processing;
Setting corresponding data mapping and format conversion strategies for different SaaS application components according to the unified data model;
According to the data mapping and format conversion strategy, performing data mapping and format conversion on the corresponding preprocessed historical monitoring data to obtain isomorphic multiple standard historical monitoring data;
According to a plurality of standard historical monitoring data, a data analysis model is constructed by using a deep learning algorithm, and a plurality of historical data analysis results are generated;
According to a plurality of historical data analysis results, a resource scheduling model is constructed by using a group intelligent optimizing algorithm, and a plurality of historical resource scheduling schemes are generated;
and constructing an upgrade strategy generation model by using a reinforcement learning algorithm according to a plurality of historical data analysis results and corresponding historical resource scheduling schemes, and generating a plurality of historical upgrade strategy generation experiences.
Further, a data analysis model is built based on BiLSTM-Attention-RF algorithm, and the data analysis model comprises an input layer, an Attention weight layer built based on an Attention mechanism, a data feature extraction layer built based on BiLSTM algorithm, a data analysis prediction layer built based on RF algorithm and an output layer which are connected in sequence;
The resource scheduling model is constructed based on ICPO algorithm, and comprises an optimizing target generating module, an optimal solution searching module and a scheme generating module which are sequentially connected;
an upgrade policy generation model is built based on the DQN algorithm and includes an agent, a deep Q network, and an empirical playback pool.
Further, collecting real-time monitoring data of the SaaS application component, and carrying out standardized processing on the real-time monitoring data to obtain standard real-time monitoring data of the SaaS application component, wherein the method comprises the following steps:
collecting real-time monitoring data of a SaaS application component in the SaaS platform, and preprocessing the real-time monitoring data to obtain preprocessed real-time monitoring data;
And according to the data mapping and format conversion strategy of the SaaS application component, performing data mapping and format conversion on the preprocessed real-time monitoring data to obtain the standard real-time monitoring data of the SaaS application component.
Further, according to standard real-time monitoring data, using a data analysis model to perform data analysis to obtain a real-time data analysis result of the SaaS application component, including the following steps:
performing sequence conversion on standard real-time monitoring data to obtain an input data sequence, and inputting the input data sequence into an input layer of a data analysis model;
Weighting the input data sequence according to a preset attention weight value 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;
extracting a weighted data feature sequence of the input data sequence by using a data feature extraction layer, converting the data feature sequence into a data feature vector, and inputting the data feature vector into a data analysis prediction layer;
extracting a plurality of key data features of the data feature vector by using a data analysis prediction layer, and carrying out data analysis prediction according to the plurality of key data features to obtain a data analysis prediction code;
And decoding the data analysis prediction codes by using an output layer to obtain a real-time data analysis result of the SaaS application component.
Further, according to the real-time data analysis result, using a resource scheduling model to perform resource scheduling to obtain a real-time resource scheduling scheme of the SaaS application component, including the following steps:
Analyzing the real-time data analysis result by using an optimizing target generating module of the resource scheduling model to obtain a real-time optimizing target and a real-time searching space, and generating a real-time target function according to the real-time optimizing target;
Using an optimal solution searching module to perform iterative optimization in a real-time searching space according to a real-time objective function to obtain a position coding vector corresponding to an optimal solution;
And decoding the position coding vector corresponding to the optimal solution by using a scheme generating module to obtain a real-time resource scheduling scheme of the SaaS application component.
Further, according to the real-time data analysis result and the real-time resource scheduling scheme, an upgrade strategy generation model is used for generating an upgrade strategy to obtain a real-time upgrade strategy of the SaaS application component, and the method comprises the following steps:
searching in an experience playback pool of an upgrade strategy generation model according to a real-time data analysis result and a real-time resource scheduling scheme, and extracting a plurality of matching history upgrade strategy generation experiences;
updating the action space of the updating strategy generation model according to a plurality of experience generated by matching the history updating strategy to obtain updated action spaces of a plurality of possible actions;
updating the state space of the upgrade strategy generation model according to the real-time data analysis result and the real-time resource scheduling scheme to obtain an updated state space containing a plurality of states;
Inputting the updated state space into a deep Q network, controlling the deep Q network by using an intelligent agent, and outputting a Q value sequence of a possible action sequence in the updated action space;
according to a preset reward function, carrying out iterative updating on the Q value sequence to obtain an updated Q value sequence until the iterative times reach a time threshold;
according to the greedy strategy, taking the possible action of the highest updated Q value in the updated Q value sequence of each iteration as the execution action of the corresponding state;
and integrating the execution actions of all states in the updated state space to obtain the real-time upgrading strategy of the SaaS application component.
Further, the real-time data analysis results comprise real-time application component state analysis results, real-time upgrading requirement analysis results and real-time upgrading risk analysis results;
the real-time application component state analysis results comprise a real-time cost analysis result, a real-time performance analysis result, a real-time resource utilization rate analysis result and a real-time stability analysis result;
the real-time resource scheduling scheme comprises a real-time resource demand condition and a real-time resource scheduling strategy;
The real-time upgrade strategy comprises a real-time upgrade decision and a real-time resource scheduling decision.
The system comprises a model building unit, a data acquisition unit, a data analysis unit, a resource scheduling unit, an upgrading strategy generation unit and an offline upgrading unit which are sequentially connected, wherein the SaaS platform is provided with a plurality of SaaS application components, and each SaaS application component is respectively connected with the data acquisition unit and the offline upgrading unit.
The beneficial effects of the invention are as follows:
The invention discloses a method and a system for upgrading a SaaS application component, which are used for realizing the intellectualization of an upgrading process by constructing a data analysis model, a resource scheduling model and an upgrading strategy generation model, reducing manual intervention, lowering cost investment, enhancing practicability and improving upgrading efficiency and accuracy; the data analysis model has the capability of monitoring the state of the application component in real time, carries out real-time analysis, can discover and solve the problem in time, reduces the upgrading risk, the resource scheduling model monitors the service condition of resources, intelligently schedules the resources, improves the resource utilization rate, reduces the operation cost, can realize the optimized allocation of the resources in the upgrading process, furthest reduces service interruption, ensures the user experience and service continuity, and the upgrading strategy generation model carries out upgrading strategy generation according to the data analysis result and the resource scheduling scheme, so that the upgrading process can adapt to the current state and the requirement of the application component, improves the flexibility and the adaptability of the strategy, realizes the dynamic generation of the upgrading strategy, can be adjusted according to the real-time system state and the environment change, and ensures that the upgrading strategy is always matched with the actual requirement.
Other advantageous effects of the present invention will be further described in the detailed description.
Drawings
FIG. 1 is a flow chart of a method for upgrading SaaS application components in the invention.
FIG. 2 is a block diagram of the SaaS application component upgrade system of the present invention.
Detailed Description
The invention is further illustrated by the following description of specific embodiments in conjunction with the accompanying drawings.
Example 1:
as shown in fig. 1, the embodiment provides a method for upgrading a SaaS application component, which includes the following steps:
S1, in a SaaS platform, constructing a data analysis model, a resource scheduling model and an upgrade strategy generation model according to historical monitoring data of different SaaS application components, wherein the method comprises the following steps of:
S1-1, in the SaaS platform, constructing a unified data model according to a universal data format supported by the SaaS platform, wherein the method comprises the following steps of:
S1-1-1, setting a plurality of core data elements in a SaaS platform according to a general data format supported by the SaaS platform;
s1-1-2, setting a data structure of a unified data model, such as a table, a field, a relation and the like, according to a plurality of core data elements;
S1-1-3, setting data relations among core data elements of a unified data model according to a plurality of core data elements, wherein the data relations comprise one-to-one, one-to-many, many-to-many and the like;
S1-1-4, setting data constraint of each core data element according to a data structure of a plurality of core data elements and a unified data model and a data relation among a plurality of core data elements, such as data types, value ranges, default values and the like;
S1-1-5, constructing a corresponding unified data model according to a plurality of core data elements, a data structure of the unified data model, a data relation among the plurality of core data elements and a data constraint of each core data element;
S1-2, collecting a plurality of historical monitoring data of different SaaS application components, and preprocessing the heterogeneous historical monitoring data to obtain historical monitoring data after intervention processing;
Preprocessing comprises sequentially performing data cleaning, repeated elimination, normalization and data dimension reduction, improving the data quality, scaling the data to a uniform proportion so as to improve the speed and effect of model training, performing data dimension reduction by using principal component analysis (PRINCIPAL COMPONENT ANALYSIS, PCA), and transforming the original data into a group of linearly independent representations of each dimension through linear transformation for removing noise and redundancy in the data, so that the data is easier to understand and process;
The historical monitoring data comprises historical online running data, historical offline log data and historical basic attribute data of the SaaS application component;
The online operation data comprise operation flow data, resource use condition, operation memory resource use data, operation network resource use data, operation cache resource use condition and the like of the SaaS application component, the offline log data comprise operation log data, safety monitoring and protection log, login log and the like of the SaaS application component, and the basic attribute data comprise component version information, performance index information, compatibility information and the like;
The SaaS application components of the SaaS platform typically include user interface components (dashboards, configuration panels, interactive tools), business logic components (business rules 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), and the like;
s1-3, setting corresponding data mapping and format conversion strategies for different SaaS application components according to the unified data model;
S1-4, performing data mapping and format conversion on the corresponding preprocessed historical monitoring data according to a data mapping and format conversion strategy to obtain isomorphic multiple standard historical monitoring data;
The data mapping and converting strategy comprises a mapping strategy and a data converting strategy, wherein the mapping strategy is used for determining how to map heterogeneous data into a unified data model, the data converting strategy is used for executing a preset converting algorithm or script to convert the heterogeneous data into data in the unified data model according to the mapping strategy, and the data mapping and converting strategy is used for processing data isomerism and solving the inconsistency in aspects of different data formats, data types, data semantics and the like;
s1-5, constructing a data analysis model by using a deep learning algorithm according to a plurality of standard historical monitoring data, and generating a plurality of historical data analysis results;
The data analysis model is built based on a bidirectional long-short-Term Memory network (Bidirectional Long Short-Term Memory, biLSTM) -Attention-Random Forest (RF) algorithm, and comprises an input layer, an Attention weight layer built based on an Attention mechanism, a data feature extraction layer built based on a BiLSTM algorithm, a data analysis prediction layer built based on an RF algorithm and an output layer which are sequentially connected;
The historical data analysis results comprise historical application component state analysis results, historical upgrading requirement analysis results and historical upgrading risk analysis results;
the historical application component state analysis results comprise historical cost analysis results, historical performance analysis results (such as response time and scores of throughput), historical resource utilization analysis results (such as CPU (Central processing Unit), memory, storage and network resource utilization) and historical stability analysis results;
S1-6, constructing a resource scheduling model by using a group intelligent optimizing algorithm according to a plurality of historical data analysis results, and generating a plurality of historical resource scheduling schemes;
The resource scheduling model is constructed based on an improved coronaries optimizer (Improved Crested Porcupine Optimizer, ICPO) algorithm, and comprises an optimizing target generating module, an optimal solution searching module and a scheme generating module which are sequentially connected;
the historical resource scheduling scheme comprises a historical resource demand condition and a historical resource scheduling strategy;
S1-7, constructing an upgrade strategy generation model by using a reinforcement learning algorithm according to a plurality of historical data analysis results and corresponding historical resource scheduling schemes, and generating a plurality of historical upgrade strategy generation experiences, wherein the method comprises the following steps of:
S1-7-1, constructing an intelligent body and an experience playback pool by taking a SaaS application component upgrade strategy generation problem as a simulation environment of a DQN algorithm;
S1-7-2, defining an action space of the DQN algorithm according to the analysis result of the historical data and a corresponding historical resource scheduling scheme, wherein the action space comprises state parameters corresponding to the analysis result of the historical data and state parameters corresponding to the historical resource scheduling scheme;
S1-7-3, defining an action space of the DQN algorithm according to actions required to be output by an upgrade strategy, wherein the action space comprises an upgrade action of a SaaS application component and a resource scheduling action;
s1-7-4, defining a reward function of the DQN algorithm according to the possible influence condition of each action in the action space, and evaluating the advantages and disadvantages or the influence of the actions;
s1-7-5, constructing an input layer, a plurality of hidden layers and an output layer of a depth Q network, connecting the input layer to a state space, and connecting the output layer to an action space;
s1-7-6, based on a state space, an action space and a reward function, optimizing training is carried out on a deep Q network and an agent according to a plurality of historical data analysis results and a corresponding historical resource scheduling scheme, an upgrade strategy generation model is constructed, and a plurality of generated historical upgrade strategy generation experiences are generated;
S1-7-7, generating experiences by a plurality of history upgrading strategies and storing the experiences into an experience playback pool;
the upgrade strategy generation model is constructed based on a Deep Q Network (DQN) algorithm, and comprises an agent, a Deep Q Network and an experience playback pool;
S2, collecting real-time monitoring data of the SaaS application assembly, and carrying out standardized processing on the real-time monitoring data to obtain standard real-time monitoring data of the SaaS application assembly, wherein the method comprises the following steps:
S2-1, acquiring real-time monitoring data of a SaaS application component in a SaaS platform, and preprocessing the real-time monitoring data to obtain preprocessed real-time monitoring data;
the real-time monitoring data comprise real-time online running data, real-time offline log data and real-time basic attribute data of the SaaS application component;
the preprocessing comprises normalization and data dimension reduction which are sequentially carried out;
S2-2, performing data mapping and format conversion on the preprocessed real-time monitoring data according to a data mapping and format conversion strategy of the SaaS application component to obtain standard real-time monitoring data of the SaaS application component;
S3, carrying out data analysis by using a data analysis model according to standard real-time monitoring data to obtain a real-time data analysis result of the SaaS application component, wherein the method comprises the following steps of:
s3-1, carrying out sequence conversion on standard real-time monitoring data to obtain an input data sequence, and inputting the input data sequence into an input layer of a data analysis model;
s3-2, weighting the input data sequence according to a preset attention weight value 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;
s3-3, extracting a weighted data feature sequence of the input data sequence by using a data feature extraction layer, converting the data feature sequence into a data feature vector, and inputting the data feature vector into a data analysis prediction layer;
S3-4, extracting a plurality of key data features of the data feature vector by using a data analysis prediction layer, and carrying out data analysis prediction according to the plurality of key data features to obtain a data analysis prediction code, wherein the method comprises the following steps of:
s3-4-1, extracting feature contribution degrees of a plurality of alternative data features in the data feature vector by using the RF structure trained by the data analysis prediction layer;
The formula is:
in the formula, Feature contribution degrees for j' th candidate data features; The characteristic contribution degree of the j 'alternative data characteristic in the i decision tree of the random forest is provided, i is a decision tree indication quantity, j' is an alternative data characteristic indication quantity, and n is a decision total number;
Wherein GI m、GIl、GIr" is the base index of decision tree node m, node l and node r 'of random forest, p mk" is the proportion of class K' in decision tree node m, K is the total number of classes, m, l and r 'are the node indication quantity, K' is the class indication quantity;
S3-4-2, carrying out normalization processing on the feature contribution degrees of the plurality of candidate data features to obtain a plurality of corresponding normalized feature contribution degrees;
The formula is:
Wherein VIM j' is the feature contribution degree after normalization processing, J is the total number of the candidate data features;
s3-4-3, generating a feature selection standard value of a plurality of alternative data features according to the normalized feature contribution degree;
The formula is:
In the formula, CFC j' is a feature selection standard value of a j ' alternative data feature, VIM j" is a feature contribution degree after normalization processing of the j ' alternative data feature, and j ' is an alternative data feature indication quantity;
S3-4-4, performing power-down sorting on the candidate data features according to the feature selection standard value, and selecting the first M candidate data features as key data features to obtain M key data features, wherein M is the total number of the key data features;
s3-4-5, carrying out data analysis and prediction according to M key data characteristics to obtain data analysis and prediction codes;
S3-5, decoding the data analysis prediction codes by using an output layer to obtain real-time data analysis results of the SaaS application assembly;
The real-time data analysis result comprises a real-time application component state analysis result, a real-time upgrading requirement analysis result and a real-time upgrading risk analysis result;
the real-time application component state analysis results comprise real-time cost analysis results, real-time performance analysis results (such as response time and throughput scoring), real-time resource utilization analysis results (such as CPU (Central processing Unit), memory, storage and network resource utilization) and real-time stability analysis results;
s4, carrying out resource scheduling by using a resource scheduling model according to the real-time data analysis result to obtain a real-time resource scheduling scheme of the SaaS application component, wherein the method comprises the following steps of:
s4-1, analyzing a real-time data analysis result by using an optimizing target generation module of a resource scheduling model to obtain a real-time optimizing target and a real-time searching space, and generating a real-time target function according to the real-time optimizing target;
In this embodiment, if the real-time resource utilization rate analysis result in the real-time data analysis result is that the utilization rate is low, resource scheduling is required to improve the abnormal situation of the SaaS application component with low utilization rate;
the maximum resource utilization rate is taken as a real-time optimizing target, and a formula of the constructed real-time objective function is as follows:
wherein F is a real-time objective function, ζ (J) is the utilization rate of the jth resource, including CPU, memory, storage and network resources, J is a resource indication quantity, J is the total number of resources;
S4-2, using an optimal solution searching module, performing iterative optimization in a real-time searching space according to a real-time objective function to obtain a position coding vector corresponding to an optimal solution, and comprising the following steps:
S4-2-1, using an optimal solution searching module to encode a resource scheduling scheme into ICPO individual position encoding vectors in ICPO algorithm, and using a real-time objective function as a ICPO individual fitness function;
s4-2-2, setting ICPO population parameters and maximum iteration times of ICPO algorithm;
S4-2-3, initializing by using a Circle chaotic mapping sequence according to the position coding vector of ICPO individuals and ICPO population parameters, and generating a plurality of initial ICPO individuals of an initial ICPO population;
The formula is:
in the formula, An initial ICPO individual, namely an initial solution, of the Circle chaotic map; initial ICPO individuals generated randomly; i' is ICPO individual indicative;
S4-2-4, introducing a circulating population reduction mechanism, and limiting the number of individuals in the ICPO population parameters to obtain updated ICPO population parameters of the next iteration;
The formula is:
Wherein S t+1 is the number of individuals in ICPO population parameters of the (t+1) th iteration, S t is the number of individuals in ICPO population parameters of the (t) th iteration, S min is the minimum number of individuals in ICPO population parameters, a' is a function evaluation parameter, V is a function evaluation cycle parameter, V max is a maximum function evaluation cycle parameter, and t is an iteration number indication quantity;
s4-2-5, calculating initial fitness values of initial ICPO individuals in the initial ICPO population according to the fitness function;
s4-2-6, updating an initial ICPO population to obtain an updated ICPO population by using a first defense strategy, a second defense strategy, a third defense strategy and a fourth defense strategy according to the initial fitness value and the updated ICPO population parameters;
The formula of the first defense strategy is:
in the formula, ICPO individuals updated for the first defensive range; Is the initial ICPO individuals in the first defense range, wherein tau 1 is a random number based on normal distribution, and tau 2 is a random value in the interval [0,1 ]; is the optimal solution in the first defense range; vector generated between true optimal solution and optimal solution randomly selected from ICPO groups in the first defense range, i' is ICPO individual indication quantity, t is iteration indication quantity;
The formula of the second defense strategy is:
in the formula, ICPO individuals updated for the second defensive range; ICPO individuals initially within a second defensive range; τ 3 is a random value in interval [0,1 ]; The two random integers are respectively the r1 and r2 initial ICPO individuals, wherein r1 and r2 are two random integers between [1 and S ]; a vector generated between a true optimal solution and an optimal solution randomly selected from ICPO populations within a second defensive range;
The formula of the third defense strategy is:
in the formula, ICPO individuals updated for a third defensive range; ICPO individuals initially within a third defensive range; searching an upper limit vector for a third defense range; The number of the original ICPO units is r2 and r3 respectively, wherein r3 is a random integer between [1, S ]; an odor diffusion factor defined for the fitness function; lambda t is a defense factor; Searching for a direction control parameter;
the formula of the fourth defense strategy is:
in the formula, ICPO individuals updated in the fourth defense range; ICPO individuals initially within a fourth defensive range; Is the optimal solution in the fourth defense range, τ 4、τ5 is the random value in interval [0,1], and λ t is the defense factor; Searching for a direction control parameter; A' is a convergence speed factor;
S4-2-7, performing dynamic reverse learning on the updated ICPO population by using a dynamic reverse learning algorithm to generate a dynamic reverse ICPO population;
The formula is:
in the formula, The vector is a ICPO body with dynamic reversal, gamma' is a decreasing inertia coefficient, and L max、Lmin is a maximum value and a small value of vector space respectively; ICPO individuals who are updated;
S4-2-8, calculating the fitness value of all ICPO individuals in the updated ICPO population and the dynamically reversed ICPO population according to the fitness function, taking ICPO individuals with the largest fitness value as optimal individuals, and reserving the optimal individuals;
S4-2-9, outputting an optimal solution corresponding to the optimal individual if the iteration times of the iterative optimization reach the maximum iteration times or the fitness value of the optimal individual meets the requirement, and obtaining a position coding vector corresponding to the optimal solution;
S4-3, decoding the position coding vector corresponding to the optimal solution by using a scheme generation module to obtain a real-time resource scheduling scheme of the SaaS application component;
the real-time resource scheduling scheme comprises a real-time resource demand condition and a real-time resource scheduling strategy;
S5, generating an upgrade strategy by using an upgrade strategy generation model according to a real-time data analysis result and a real-time resource scheduling scheme to obtain a real-time upgrade strategy of the SaaS application component, wherein the method comprises the following steps of:
s5-1, searching in an experience playback pool of an upgrade strategy generation model according to a real-time data analysis result and a real-time resource scheduling scheme, and extracting a plurality of matching history upgrade strategy generation experiences;
S5-2, generating experiences according to a plurality of matching history upgrading strategies, and updating an action space of an upgrading strategy generation model to obtain an updated action space A ' = [ a ' 1,...,a'j",...,a'I ] of a plurality of possible actions, wherein a ' j" is an updated j ' action value, j ' is an action indication quantity, and I is the total number of dimensions of the action space;
In this embodiment, the upgrading actions of the SaaS application component in the action space include non-upgrading, small version upgrading, large version upgrading, partial upgrading, staged upgrading, upgrading after resource optimization, rollback to the last version, small version upgrading after more CPU resources are allocated, large version upgrading after more memory resources are allocated, large version upgrading after resources are dynamically extended, and the resource scheduling includes resource allocation, resource release, load balancing, resource reservation, dynamic extension, priority scheduling, and the like;
S5-3, updating a state space of an upgrade strategy generation model according to a real-time data analysis result and a real-time resource scheduling scheme to obtain an updated state space S '= [ S' 1,...,s'i",...,s'I' ] containing a plurality of states, wherein S 'i" is an updated I' state value, I 'is a state indication quantity, and I' is the total number of state space dimensions;
In this embodiment, the state parameters corresponding to the real-time data analysis result in the action space include a real-time cost analysis score, a real-time performance analysis score, a real-time resource utilization rate, a real-time stability analysis score, a real-time upgrade requirement, a real-time upgrade risk analysis score, and the like, and the state parameters corresponding to the real-time resource scheduling scheme include a real-time resource requirement and a real-time resource scheduling policy;
S5-4, inputting the updated state space S '= [ S' 1,...,s'i',...,s'I ] into a depth Q network, controlling the depth Q network by using an intelligent agent, and outputting a Q value sequence of a possible action sequence in the updated action space A '= [ a' 1,...,a'j',...,a'I ];
S5-5, carrying out iterative updating on the Q value sequence according to a preset reward function to obtain an updated Q value sequence until the iterative times reach a time threshold;
The formula is:
Q(s'p',a'p')＝(1-α")·Q(sp',ap')+α"·(R(sp',ap',s'p')+γ·Qmax(sp',ap'))
Wherein Q (s ' p',a'p') is an updated Q value corresponding to an updated state value s ' p' and an updated action value a ' p', Q (s p',ap') is a predicted Q value corresponding to a state value s p' and an action value a p', alpha ' is a learning rate, Q max(sp',ap') is the highest predicted Q value, p ' is a comprehensive instruction amount, and gamma is an update parameter;
S5-6, taking possible actions of the highest updated Q value in the updated Q value sequence of each iteration as executing actions of the corresponding state according to a greedy strategy;
S5-7, integrating execution actions of all states in the updated state space to obtain a real-time upgrading strategy of the SaaS application component;
the real-time upgrading strategy comprises a real-time upgrading decision and a real-time resource scheduling decision;
in this embodiment, the real-time upgrade policy includes minor version upgrade after more CPU resources are allocated, CPU resource allocation is performed by the current SaaS application component, and resource release is performed by other SaaS application components;
and S6, carrying out offline upgrading on the corresponding SaaS application assembly according to the real-time upgrading strategy to obtain an upgraded SaaS application assembly, and returning to the monitoring data acquisition step.
Example 2:
As shown in fig. 2, the present embodiment provides a SaaS application component upgrading system, configured to implement a SaaS application component upgrading method, where the system is disposed on a SaaS platform, and the system includes a model building unit, a data collecting unit, a data analysis unit, a resource scheduling unit, an upgrade policy generating unit, and an offline upgrading unit that are sequentially connected, where the SaaS platform is provided with a plurality of SaaS application components, and each SaaS application component is connected to the data collecting unit and the offline upgrading unit respectively;
The model building unit is used for building a data analysis model, a resource scheduling model and an upgrade strategy generation model according to the historical monitoring data of different SaaS application components in the SaaS platform;
The data acquisition unit is used for acquiring real-time monitoring data of the SaaS application component and carrying out standardized processing on the real-time monitoring data to obtain standard real-time monitoring data of the SaaS application component;
the data analysis unit is used for monitoring data in real time according to the standard, and performing data analysis by using a data analysis model to obtain a real-time data analysis result of the SaaS application component;
the resource scheduling unit is used for performing resource scheduling by using a resource scheduling model according to the real-time data analysis result to obtain a real-time resource scheduling scheme of the SaaS application component;
The upgrade strategy generation unit is used for generating an upgrade strategy by using an upgrade strategy generation model according to the real-time data analysis result and the real-time resource scheduling scheme to obtain a real-time upgrade strategy of the SaaS application component;
And the offline upgrading unit is used for carrying out offline upgrading on the corresponding SaaS application components according to the real-time upgrading strategy to obtain the upgraded SaaS application components.
The invention discloses a method and a system for upgrading a SaaS application component, which are used for realizing the intellectualization of an upgrading process by constructing a data analysis model, a resource scheduling model and an upgrading strategy generation model, reducing manual intervention, lowering cost investment, enhancing practicability and improving upgrading efficiency and accuracy; the data analysis model has the capability of monitoring the state of the application component in real time, carries out real-time analysis, can discover and solve the problem in time, reduces the upgrading risk, the resource scheduling model monitors the service condition of resources, intelligently schedules the resources, improves the resource utilization rate, reduces the operation cost, can realize the optimized allocation of the resources in the upgrading process, furthest reduces service interruption, ensures the user experience and service continuity, and the upgrading strategy generation model carries out upgrading strategy generation according to the data analysis result and the resource scheduling scheme, so that the upgrading process can adapt to the current state and the requirement of the application component, improves the flexibility and the adaptability of the strategy, realizes the dynamic generation of the upgrading strategy, can be adjusted according to the real-time system state and the environment change, and ensures that the upgrading strategy is always matched with the actual requirement.
The invention is not limited to the alternative embodiments described above, but any person may derive other various forms of products in the light of the present invention. The above detailed description should not be construed as limiting the scope of the invention, which is defined in the claims and the description may be used to interpret the claims.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411681986.3A CN119621101A (en) | 2024-11-22 | 2024-11-22 | A SaaS application component upgrade method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411681986.3A CN119621101A (en) | 2024-11-22 | 2024-11-22 | A SaaS application component upgrade method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN119621101A true CN119621101A (en) | 2025-03-14 |
Family
ID=94891616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411681986.3A Pending CN119621101A (en) | 2024-11-22 | 2024-11-22 | A SaaS application component upgrade method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119621101A (en) |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190250949A1 (en) * | 2018-02-14 | 2019-08-15 | Citrix Systems, Inc. | Auto-scaling for allocation of cloud service resources in application deployments |
| US20190312800A1 (en) * | 2015-07-27 | 2019-10-10 | Datagrid Systems, Inc. | Method, apparatus and system for real-time optimization of computer-implemented application operations using machine learning techniques |
| DE202024101468U1 (en) * | 2024-03-24 | 2024-04-23 | Shweta Chauhan | Dynamic workflow optimization system for improved task management efficiency |
| CN118012571A (en) * | 2024-02-04 | 2024-05-10 | 黄淮学院 | A container scheduling method and system based on cloud platform |
| US20240193467A1 (en) * | 2022-12-13 | 2024-06-13 | International Business Machines Corporation | Intelligent upgrade workflow for a container orchestration system |
| CN118297329A (en) * | 2024-04-08 | 2024-07-05 | 西南石油大学 | Digital twinning-based gas transmission and distribution scheduling method and system |
| CN118445036A (en) * | 2024-04-16 | 2024-08-06 | 中电万维信息技术有限责任公司 | Intelligent scheduling method for data sharing exchange task |
| CN118463337A (en) * | 2024-06-03 | 2024-08-09 | 华电四川发电有限公司宝珠寺水力发电厂 | Factory building exhaust remote control method and system based on sensor technology |
| CN118503915A (en) * | 2024-06-18 | 2024-08-16 | 浙江极象科技有限公司 | Multi-source heterogeneous data fusion method and system |
| CN118519650A (en) * | 2024-04-24 | 2024-08-20 | 天翼云科技有限公司 | GPT technology-based hybrid cloud automatic upgrading method |
| CN118535195A (en) * | 2024-07-10 | 2024-08-23 | 深圳益邦阳光有限公司 | Firmware upgrade management method, device, equipment and storage medium |
| CN118765020A (en) * | 2024-09-06 | 2024-10-11 | 四川菲尼特数字科技集团有限公司 | A smart street lamp automatic control system and method |
| CN118842656A (en) * | 2024-09-23 | 2024-10-25 | 北京大华无线电仪器有限责任公司 | Power supply test data management method and system based on machine learning |
-
2024
- 2024-11-22 CN CN202411681986.3A patent/CN119621101A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190312800A1 (en) * | 2015-07-27 | 2019-10-10 | Datagrid Systems, Inc. | Method, apparatus and system for real-time optimization of computer-implemented application operations using machine learning techniques |
| US20190250949A1 (en) * | 2018-02-14 | 2019-08-15 | Citrix Systems, Inc. | Auto-scaling for allocation of cloud service resources in application deployments |
| US20240193467A1 (en) * | 2022-12-13 | 2024-06-13 | International Business Machines Corporation | Intelligent upgrade workflow for a container orchestration system |
| CN118012571A (en) * | 2024-02-04 | 2024-05-10 | 黄淮学院 | A container scheduling method and system based on cloud platform |
| DE202024101468U1 (en) * | 2024-03-24 | 2024-04-23 | Shweta Chauhan | Dynamic workflow optimization system for improved task management efficiency |
| CN118297329A (en) * | 2024-04-08 | 2024-07-05 | 西南石油大学 | Digital twinning-based gas transmission and distribution scheduling method and system |
| CN118445036A (en) * | 2024-04-16 | 2024-08-06 | 中电万维信息技术有限责任公司 | Intelligent scheduling method for data sharing exchange task |
| CN118519650A (en) * | 2024-04-24 | 2024-08-20 | 天翼云科技有限公司 | GPT technology-based hybrid cloud automatic upgrading method |
| CN118463337A (en) * | 2024-06-03 | 2024-08-09 | 华电四川发电有限公司宝珠寺水力发电厂 | Factory building exhaust remote control method and system based on sensor technology |
| CN118503915A (en) * | 2024-06-18 | 2024-08-16 | 浙江极象科技有限公司 | Multi-source heterogeneous data fusion method and system |
| CN118535195A (en) * | 2024-07-10 | 2024-08-23 | 深圳益邦阳光有限公司 | Firmware upgrade management method, device, equipment and storage medium |
| CN118765020A (en) * | 2024-09-06 | 2024-10-11 | 四川菲尼特数字科技集团有限公司 | A smart street lamp automatic control system and method |
| CN118842656A (en) * | 2024-09-23 | 2024-10-25 | 北京大华无线电仪器有限责任公司 | Power supply test data management method and system based on machine learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116611813B (en) | Intelligent operation and maintenance management method and system based on knowledge graph | |
| CN118503915B (en) | Multi-source heterogeneous data fusion method and system | |
| CN118194487A (en) | Automatic arrangement method, medium and system for circuit and electric equipment | |
| CN117390550A (en) | Low-carbon park carbon emission dynamic prediction method and system considering emission training set | |
| CN119598443A (en) | Human resources authority management method and system | |
| CN101930566B (en) | Hydrological experimental simulation system and method based on parallel system | |
| CN113536508A (en) | A method and system for classifying manufacturing network nodes | |
| CN115775051B (en) | A household power load probability prediction method and its application | |
| CN119621101A (en) | A SaaS application component upgrade method and system | |
| CN116150189A (en) | A method and system for adaptive multi-table join cardinality estimation | |
| CN118780449B (en) | A method for predicting power load | |
| CN120016441A (en) | Power system voltage stability boundary prediction method and system based on graph neural network | |
| CN117493372B (en) | SQL statement generation method and device based on relationship perception | |
| CN120013042A (en) | A load intelligent decision-making management method based on knowledge graph | |
| CN115640896B (en) | Household user power load prediction method and related equipment in multi-user scenario | |
| Liu et al. | Automatic joint optimization of algorithm-level compression and compiler-based acceleration with reinforcement learning for dnn in edge devices | |
| CN120951238A (en) | Energy consumption prediction method and system based on multimodal large model and context-awareness | |
| CN110807599A (en) | Method, device, server and storage medium for deciding electrochemical energy storage scheme | |
| CN118411002B (en) | A method, system, device and storage medium for accessing multiple types of power loads | |
| CN119202175B (en) | Intelligent question-answering method for power issues | |
| CN120566853A (en) | Method and system for automatically reducing power consumption of intelligent inverter system | |
| Zhai et al. | The Security Verification and Analysis of Information Nodes in Cyberspace by Deep Spiking Algorithm | |
| CN121480930A (en) | Method and device for determining comfort level of underground space | |
| CN119398239A (en) | A method and system for predicting coal inventory in power plants based on artificial intelligence | |
| CN120911243A (en) | Equipment cluster simulation deduction method and system based on digital twin driving |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |