System for realizing multi-application integration based on middle platform architecture

Abstract

The invention relates to the technical field of information, in particular to a system for realizing multi-application integration based on a middle platform architecture. Comprising a middle station service module which supports cross-application service calling and data sharing by constructing a middle station service layer; the resource management module is used for optimizing resource allocation by utilizing a virtualization technology and cloud load balancing; the security control module is used for implementing unified password management and encrypted transmission to guarantee data security; the interface specification module adopts a unified butt joint mode to simplify an application integration process; and the data analysis module is used for collecting user behavior data and mining and analyzing the data. Therefore, government affair service efficiency and quality are effectively improved, and deepening development of government affair digital transformation is promoted.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a system for implementing multi-application integration based on a middle platform architecture. Background Art

[0002] With the deepening of the digital transformation of government affairs, numerous application scenarios and innovative models have emerged, such as: one-stop service, one-stop service, smart government affairs, digital convenience, etc. These innovations in models have become a new tool to promote the improvement of quality and efficiency of government services.

[0003] Although the digitalization of government affairs has made some progress, it still faces challenges of technical architecture defects and decentralized services. New technology applications and security governance are weak, the underlying architecture is not unified, resulting in inefficient data docking, closed external network data architecture, and insufficient utilization of hardware resources. At the same time, application scenarios are scattered, information islands are serious, administrative approval processes are imperfect, and there is a lack of efficient application development models and widely used public support tools. The core issue is how to build a unified, efficient, and secure government technology architecture and service system. Summary of the invention

[0004] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to provide a system for implementing multi-application integration based on a middle-office architecture, so as to solve the problem of how to build a unified, efficient and secure government technology architecture and service system.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a system for implementing multi-application integration based on a middle platform architecture, comprising: The middle office service module is used to build a middle office service layer that supports service calls and data sharing across multiple government applications; The resource management module uses virtualization technology to dynamically allocate hardware resources, encapsulate physical server resources into virtual machine instances, deploy cloud load balancers, and dynamically adjust resource allocation between virtual machine instances based on the real-time load of government applications; The security control module implements a unified user login password management strategy, formulates a user account lock strategy, sets a password validity period, and uses the AES symmetric encryption algorithm for data transmission encryption; The interface specification module adopts a unified docking method and formulates unified interface specifications, data exchange formats and business logic models; The data analysis module collects and analyzes user behavior data, embeds tracking code in the pages of government applications, records user behavior data, cleans the collected data, removes invalid and duplicate data, and mines and analyzes the cleaned data.

[0006] Preferably, in a possible implementation of the first aspect, the framework of the middle platform service layer specifically includes: The front-end presentation layer uses a combination of Vue, Flutter and WeChat API frameworks to support multi-platform display; The backend logic layer is based on the SpringCloud microservice architecture and combines the RabbitMQ message queue to achieve efficient service calls; The persistence layer uses the cluster mode of PostGreSql database combined with Redis cache and ElasticSearch search engine.

[0007] Preferably, in a possible implementation manner of the first aspect, the dynamic allocation of hardware resources is specifically implemented as follows: Build a virtualized resource pool to encapsulate physical server resources into independent virtual machine instances; Deploy a cloud load balancer to monitor the load of each virtual machine in real time and automatically adjust resource allocation using a weighted round-robin algorithm; Introduce a cache system to reduce direct database access; Set resource usage thresholds. When any resource metric exceeds or falls below the threshold, resource expansion or reduction operations are automatically triggered.

[0008] Preferably, in a possible implementation manner of the first aspect, the automatically triggered resource expansion or reduction operation is specifically: When the resource load is detected to exceed the preset upper limit, a new virtual machine instance is automatically started to expand resources; When resource usage continues to be lower than the preset lower limit, idle virtual machines are automatically shut down and resources are released.

[0009] Preferably, in a possible implementation of the first aspect, the unified password management policy is specifically implemented as follows: Integrated account lock function, setting lock conditions and automatic reset mechanism; Dynamically generate a unique AES encryption key for each integrated government application to ensure data transmission security.

[0010] Preferably, in a possible implementation manner of the first aspect, the unified docking method includes interface docking, and the specific implementation steps include: Create government applications and automatically generate government application identifiers; Add government application version and submit for review; The administrator reviews the government application version. If it passes, the version is enabled and the government application configuration is performed. After the configuration is completed, it is displayed in the APP; if it fails, the current process is terminated and the government application is recreated; Create an interface and submit it for review; The administrator reviews the interface. If the review passes, the interface is bound to the enabled government application; if it fails, the current process ends and the interface is recreated.

[0011] Preferably, in a possible implementation of the first aspect, the unified docking method further includes URL docking, and the specific implementation steps include: Create government applications; Add the link version and submit for review; The administrator reviews the link version. If it passes, the version is enabled and the government application is configured. After the configuration is completed, the link is displayed in the APP; if it fails, the current process is terminated and the government application and its link version are recreated.

[0012] Preferably, in a possible implementation manner of the first aspect, the specific steps of collecting and analyzing user behavior data include: Collect user behavior data through embedding technology on government application pages and store the data in the database; Clean the data, remove noise, fill missing values ​​and perform standardization; Analyze data and build user portraits; Verify the accuracy of user portraits and perform iterative optimization based on feedback.

[0013] Preferably, in a possible implementation of the first aspect, the step of building a user portrait further includes: Extract user basic attributes, behavior characteristics and preference characteristics, and construct feature vectors; Use clustering algorithms to group users and form a user grouping model; Build detailed user profiles for each user group, including feature descriptions, behavior patterns, and interest preferences; Generate a visual user portrait report.

[0014] The beneficial effects of the present invention are: by constructing a middle-end architecture, efficient integration and data sharing of multiple government applications are achieved, solving the problems of inconsistent technical architecture and decentralized services. Virtualization and cloud load balancing technologies are used to optimize resource utilization, unified password management and encrypted transmission to ensure data security, unified docking methods to simplify integration processes, and big data analysis modules to enhance decision-making support capabilities, significantly improving the efficiency and quality of government services and promoting the deepening development of digital transformation of government affairs. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A structural diagram of a multi-application integration system based on the middle platform architecture provided in this embodiment.

[0017] Figure 2 A software architecture diagram for implementing a multi-application integration system based on a middle-end architecture is provided in this embodiment.

[0018] Figure 3 A submission application and interface flow chart for implementing a multi-application integration system based on a middle-end architecture is provided in this embodiment.

[0019] Figure 4 A flowchart of submitting applications and linking a multi-application integration system based on a middle-stage architecture is provided in this embodiment.

[0020] Figure 5 A data mining and analysis flowchart for implementing a multi-application integration system based on a middle-end architecture is provided in this embodiment.

[0021] Explanation of the accompanying drawings: 1-middle platform service module, 2-resource management module, 3-security control module, 4-interface specification module, 5-data analysis module. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Embodiment 1: Figure 1 As shown, the present invention provides a system for implementing multi-application integration based on a middle platform architecture, including: The middle platform service module 1 is used to build the middle platform service layer, which supports service calls and data sharing across multiple government applications. The software architecture of the middle platform service layer is as follows: Figure 2 shown.

[0024] Specifically, the front-end presentation layer uses a combination of Vue, Flutter and WeChat API framework to support multi-platform display.

[0025] Use Vue.js for web application development. Vue.js is a progressive JavaScript framework for building user interfaces. It focuses on the view layer, is easy to use, and easy to integrate with third-party libraries or existing projects. Use Vue.js as the core framework, and use element ui and vant-ui to enrich the user interface. Use vite as a building tool, take advantage of its fast cold start and hot update features to speed up the development process, and use axios to process HTTP requests to achieve interaction between front-end and back-end data.

[0026] Use Flutter to develop iOS and Android applications. Flutter is an open source UI framework developed by Google for building high-performance, cross-platform mobile applications. Use cached_network_image to load and cache network images, reduce repeated downloads, speed up image loading, and improve interface fluency; use permission_handler to help applications request and manage user permissions to ensure the compliance of government application functions and user privacy; use path_provider to provide a cross-platform access to the file system path interface, which is convenient for reading and writing files to specific directories in the application sandbox.

[0027] Use the WeChat API framework to make interface calls. Use wx.getUserInfo to obtain the user's encrypted information; use wx.login to obtain the user's login credentials, exchange for session_key and openid, and implement user login and identity authentication; use wx.checkSession to check whether the user's login status has expired. If expired, you need to log in again; use SupportSoterAuthentication to integrate biometric authentication in the application to enhance the security of user authentication.

[0028] The backend logic layer is based on the SpringCloud microservice architecture and combined with the RabbitMQ message queue to achieve efficient service calls.

[0029] Use Spring Cloud as the microservice development framework. Spring Cloud uses components such as Spring CloudConfig to achieve externalization and centralized management of configuration; Spring Cloud provides load balancing, circuit breakers, intelligent routing and other functions, which work together to ensure system stability and availability under the microservice architecture; intelligent routing is implemented through Zuul and Spring Cloud Gateway gateway components to route, filter, and limit external requests, thereby improving the security and controllability of the system.

[0030] Use RabbitMQ as the message center. RabbitMQ is a high-performance, open source message broker software that supports multiple message protocols. Through RabbitMQ, microservices can publish messages to specified queues or switches without directly interacting with other services. Other services subscribe to these queues or routing keys as consumers to receive and process messages asynchronously. Use the Spring AMQP module to integrate RabbitMQ. Spring AMQP provides message confirmation and transaction support functions to ensure the reliability and consistency of message delivery.

[0031] The persistence layer uses the cluster mode of PostGreSql database combined with Redis cache and ElasticSearch search engine.

[0032] The persistence layer ORM framework is a technology used in software development to solve the mismatch between object-oriented programming languages ​​and relational databases. It enables developers to operate the database in an object-oriented way by providing a mapping relationship between objects and database tables.

[0033] Use PostGreSql cluster to improve the availability, performance and fault recovery capability of the database. In the PostGreSql cluster, there is a master node and multiple slave nodes. The master node is responsible for processing all write operations and most read operations, committing transactions and generating logs, synchronizing transaction logs to slave nodes in real time, and maintaining data consistency; the slave node is responsible for receiving and applying transaction logs sent by the master node in real time, maintaining data synchronization with the master node, processing read operations, sharing the load of the master node, improving the overall performance of the system, and taking over the role of the master node when the master node fails, and continuing to provide services, thereby ensuring the high availability of the database.

[0034] Use Redis cluster to improve data capacity, throughput and availability. In a Redis cluster, there are multiple Redis master nodes and Redis slave nodes, as well as Sentinel nodes for monitoring. The Redis master node is responsible for processing data writes and most read operations. Each master node in the cluster processes a subset of 16384 hash slots to achieve data sharding; the Redis slave node replicates the data of the corresponding master node and provides read operations to share the load of the master node. When the master node fails, it can be upgraded to a new master node to achieve failover; the Sentinel node is responsible for monitoring the health status of the Redis master node and slave nodes. When the master node fails, Sentinel will automatically promote the slave node to the master node to ensure the continuous operation of the system.

[0035] Use ElasticSearch cluster to provide efficient and scalable search and data analysis capabilities. In an ElasticSearch cluster, data is divided into multiple shards for storage and retrieval. ElasticSearch primary shards are the main storage unit for data. Each index is divided into one or more primary shards, which are readable and writable and are responsible for processing data addition, deletion, modification, and query operations; ElasticSearch slave shards are exact copies of primary shards, used to improve data availability and fault tolerance, are read-only, and are distributed on different nodes to balance the load and improve query performance.

[0036] Resource management module 2 uses virtualization technology to dynamically allocate hardware resources, encapsulate physical server resources into virtual machine instances, deploy cloud load balancers, and dynamically adjust resource allocation between virtual machine instances based on the real-time load of government applications.

[0037] Specifically, a virtualized resource pool is constructed to encapsulate physical server resources into independent virtual machine instances.

[0038] Install VMware ESXi on the server as the bare metal hypervisor. At the same time, configure the virtual network environment to ensure smooth internal communication between virtual machines and connection with the external network.

[0039] Using the advanced features of VMware ESXi, the CPU, memory, storage and network resources of the physical server are subdivided into multiple independent virtual machine instances. Each virtual machine has an independent operating system, application and user data, achieving efficient isolation and sharing of resources. ESXi optimizes memory usage through technologies such as memory page sharing, memory balloon technology and memory compression, and supports advanced features such as CPU hot plugging, snapshots and virtual machine migration to ensure stable operation and high availability of virtual machines.

[0040] According to the specific needs of each application, formulate resource allocation strategies and allocate appropriate CPU cores, memory quotas, disk space and network bandwidth to each virtual machine. In particular, for government applications that require high-performance I / O, allocate more CPU cores and memory, and consider using high-speed storage; while for front-end Web servers, focus on optimizing CPU performance and network bandwidth.

[0041] During the virtual machine deployment and configuration process, use VMware's management interface to create virtual machines one by one and install the corresponding operating system. Then, configure the network settings to ensure that the virtual machines can communicate smoothly and access the external network.

[0042] Deploy VMware vCenter Server virtualization monitoring tools to monitor various performance indicators of the virtualization environment in real time, including CPU usage, memory usage, disk I / O rate, and network bandwidth. By analyzing monitoring data, timely discover and solve performance bottlenecks and resource waste problems. Implement dynamic resource adjustment strategies, such as dynamically increasing or decreasing the number of CPU cores according to CPU load, adjusting memory quotas according to memory usage, etc., to ensure efficient use of resources and continuous and stable operation of the system.

[0043] Deploy a cloud load balancer to monitor the load of each virtual machine in real time and automatically adjust resource allocation according to preset policies.

[0044] In order to achieve efficient operation of cloud computing services, a cloud load balancing mechanism based on a weighted round-robin algorithm is deployed. In this embodiment, AWS Elastic Load Balancer is selected as the cloud load balancer, and the creation, installation and initialization settings of the load balancer are completed in the cloud management console. This includes configuring the basic parameters of the load balancer, such as the IP address, port number, and protocol type, to ensure that it can correctly receive and process requests from the client.

[0045] Inside the load balancer, a dedicated backend server group is created to manage all virtual machine instances participating in load balancing. Each virtual machine instance is assigned an initial weight value based on its hardware configuration, processing power, current load status, and other factors. The weight value directly determines the probability of the instance receiving new requests. The higher the weight, the greater the chance of receiving requests.

[0046] To ensure high availability of the service, a health check mechanism is implemented. Health check requests are sent to backend instances regularly, and the response status code is checked. Once an instance is found to have timed out, returns an error status code, or fails to establish a connection, the instance is considered unhealthy and removed from the distribution list to prevent requests from being forwarded to the faulty instance. At the same time, an automatic recovery mechanism is set up so that when an unhealthy instance returns to normal, it can automatically rejoin the distribution list.

[0047] The weighted polling algorithm is used as the request distribution strategy. According to the current weight value of the backend instance, the next instance to receive the request is intelligently selected. In a request cycle, instances are selected in turn in order from high to low weight values ​​to ensure that requests can be distributed to each instance according to the preset ratio. At the same time, cloud monitoring services are integrated to collect key performance indicators such as CPU usage, memory usage, network bandwidth, response time, etc. of the load balancer and backend instances in real time. Based on these performance indicators, the weight value of each instance is dynamically adjusted to cope with sudden traffic, uneven load, etc., to achieve intelligent resource allocation and load balancing optimization.

[0048] Introduce a cache system to reduce direct database access.

[0049] Install Redis on the server and configure its basic parameters, including maximum memory usage, persistence strategy, etc. Use the Redis Desktop Manager management tool for monitoring and management.

[0050] In the SpringCloud microservice architecture, when the application service layer needs to read data, it first checks whether the required data exists in the Redis cache. If it exists in the cache, it is directly read from the cache and returned to the requester, reducing database access. When the data is updated, the data in the Redis cache is also updated while the database is updated to ensure data consistency. Set a reasonable expiration time for the cached data. When the data expires, the data will be reloaded from the database to the cache when it is accessed again. At the same time, the cache can also be actively invalidated through external events.

[0051] Use Redis monitoring tools to monitor key indicators such as cache hit rate and memory usage. Adjust cache strategies based on monitoring data, such as adjusting cache size and optimizing cache data structure. Clean up cache data regularly, delete useless or expired data, and free up cache space.

[0052] Set resource usage thresholds. When any resource metric exceeds or falls below the threshold, resource expansion or reduction operations are automatically triggered.

[0053] When it is detected that the resource load exceeds the preset upper limit, a new virtual machine instance is automatically started to expand resources; when the resource utilization rate continues to be lower than the preset lower limit, the idle virtual machine is automatically shut down and the resources are released.

[0054] In this embodiment, the upper and lower limits of CPU usage are set through the system interface. The upper limit is set to 80% and the lower limit is set to 20%. This means that when the CPU usage of any virtual machine exceeds 80% continuously, the system will judge it as insufficient resources; and when the CPU usage is continuously below 20%, it is considered as idle resources.

[0055] When the CPU usage of a virtual machine is monitored to exceed 80% for two consecutive monitoring cycles (i.e., two minutes), the system decision logic module determines that the virtual machine resources are insufficient and triggers the expansion operation. The execution module then automatically starts a new virtual machine instance through the cloud management platform and configures the corresponding resources according to business needs. After the new instance is started, the system may redirect part of the traffic or tasks to the new instance through the load balancer to reduce the load on the original instance.

[0056] On the contrary, when the system detects that the CPU usage of a virtual machine is less than 20% for four consecutive monitoring cycles (i.e., four minutes), the decision logic module determines that the virtual machine resources are too idle. The execution module then analyzes the load history and current business needs of the virtual machine. If it is confirmed that the resources can be reduced, the virtual machine is automatically shut down and the computing resources it occupies are released. Before shutting down, the system ensures that all running services or applications have been properly handled, such as migrating to other instances or shutting down safely.

[0057] The system also has a self-optimization mechanism. After each expansion or reduction operation, the system will continue to monitor resource usage and make fine adjustments based on the actual results. For example, if it is found that the resource utilization of the new instance after expansion is still low, the system will reallocate resources; if it is found that the response time of some services has increased after reduction, the system will restore some of the shut down virtual machine instances.

[0058] Security control module 3 implements a unified user login password management strategy, formulates a user account lock strategy, sets a password validity period, and uses the AES symmetric encryption algorithm for data transmission encryption.

[0059] Specifically, the specific conditions for account locking are configured through the interface or API interface. When the number of consecutive login failures reaches 5, the system automatically locks the account for 30 minutes. At the same time, the automatic reset mechanism is configured, and the time interval is set to 10 minutes. After this time interval, the number of failed login attempts of the user is automatically reset to avoid users being locked for a long time due to occasional input errors.

[0060] To ensure the security of data transmission, a unique AES encryption key is dynamically generated for each integrated application. This is done automatically when the application first connects to the system. The generated key will be securely stored in the system's key management system and protected by multi-layer encryption and strict access control mechanisms to prevent unauthorized access.

[0061] The system's built-in password policy configuration function should provide a user-friendly interface or API interface, allowing administrators to configure the duration of account lockout, the upper limit of the number of failed login attempts allowed, the interval for automatically resetting the failed login count, and the password validity period according to actual needs. These configuration information will be stored in the system's configuration database and verified and executed during operations such as user login and password change.

[0062] During data transmission, the system automatically detects and identifies sensitive information (including user passwords, personal information, etc.), and encrypts this information using the AES encryption key. Whether it is an interface parameter or a return value, as long as it contains sensitive information, it should be encrypted into ciphertext before leaving the source system, and the recipient will use the corresponding key to decrypt the data after receiving it to restore the original information. This two-way encryption mechanism will effectively resist the risk of data being stolen or tampered with during transmission.

[0063] In addition, the system also has security auditing and monitoring functions. All encryption and decryption operations are recorded and corresponding logs are generated to facilitate security auditing and troubleshooting when necessary. At the same time, the system monitors the encryption status, key usage, and any abnormal behavior or potential security risks during data transmission in real time, and immediately triggers an alarm mechanism when an abnormal situation is found, notifying the administrator to take timely countermeasures.

[0064] Interface specification module 4 adopts a unified docking method and formulates unified interface specifications, data exchange formats and business logic models.

[0065] Specifically, in terms of interface specifications, we follow the RESTful API standard to design interfaces and formulate interface design guidelines, including key elements such as request methods, request headers, and response codes, to ensure that all interfaces follow unified standards. At the same time, we use the Swagger tool to automatically generate interface documents to provide third-party developers with clear and easy-to-understand access guides. In addition, we have implemented an interface version control strategy to cope with possible changes and expansion requirements.

[0066] JSON or XML is a standard format for data exchange. Both formats have wide support and good readability. In this embodiment, JSON format is used as the data exchange format. In order to ensure the consistency and accuracy of data, data structure, field type, required items and constraints are defined. At the same time, a data verification mechanism is introduced to pre-verify data at the interface layer to reduce errors and exceptions in subsequent processing.

[0067] In the design of the business logic model, the principle of domain-driven design is adopted to divide the complex business logic into clear domain models and services. This design method helps to reduce the coupling of the system and improve the maintainability and scalability of the code. At the same time, the microservice architecture is also adopted to split different business functions into independent service units to achieve more flexible and efficient deployment and management.

[0068] In order to simplify the access process of third-party applications, a graphical interface management platform is used. The platform supports developers to easily register, configure and manage interfaces, and distinguish the interface call permissions of different applications through unique identifiers. This embodiment uses [the first letter of the Chinese pinyin of the department abbreviation] + [the first letter of the Chinese pinyin of the interface name] as the unique interface identifier. In addition, an interface call monitoring function is also provided, which displays key indicators such as the request volume, response time, and error rate of the interface in real time to help timely discover and solve potential problems.

[0069] In the submission process of third-party applications and interfaces, first create the application and generate a unique application identifier. Then, submit the application version and interface information. After review, the version can be enabled and the interface can be created. During the interface creation process, each interface is assigned a unique interface identifier, and a detailed interface description and example are provided. After the interface is created, perform configuration work such as interface binding and permission setting, and submit it for review again. After all reviews are passed, the interface will be officially bound to the corresponding application for user access and use. The process of submitting applications and interfaces is as follows: Figure 3 shown.

[0070] For link processes that require quick and seamless integration, developers can achieve integration by simply configuring the URL link of the application homepage on the platform. The system will conduct basic format and content review of the link to ensure the legitimacy and validity of the link. After the review is passed, the developer can further perform configuration work such as permission setting and interface docking. Finally, the system will review the completeness of the configuration and the compatibility of the application. If passed, the application will be automatically displayed on the platform for user access. Submit application and link process as follows Figure 4 shown.

[0071] Data analysis module 5 collects and analyzes user behavior data, embeds tracking code in the pages of government applications, records user behavior data, cleans the collected data, removes invalid and duplicate data, and mines and analyzes the cleaned data. Figure 5 shown.

[0072] Specifically, front-end tracking codes are deployed on various key pages (such as homepage, service guide, online processing, consultation feedback, etc.). These codes can automatically record key behavioral data such as user click events, browsing paths, dwell time, and operation frequency. At the same time, in order to ensure the comprehensiveness and accuracy of the data, tracking is performed on key business processes (including application submission, approval progress query, etc.) to capture the user's complete service experience process.

[0073] The collected raw data will first be imported into the data processing tool for cleaning. This embodiment uses Python's Pandas library. By writing scripts, the noise in the data is identified and removed, such as outliers (including extremely short residence time, frequent invalid clicks, etc.), duplicate records, etc. For missing values, the average value is used for prediction and filling. In addition, the data is standardized to ensure the consistency of time format, data format, etc., which is convenient for subsequent analysis.

[0074] After data cleaning, statistical tools and visualization methods are used to conduct in-depth analysis of the data. Descriptive statistical analysis is performed using Python's NumPy, SciPy and other libraries to calculate the mean, median, standard deviation and other indicators of user behavior to understand the basic situation of user behavior. At the same time, correlation analysis is used to explore the degree of correlation between different behavioral indicators to provide a basis for subsequent cluster analysis. In order to more intuitively display the results of data analysis, Matplotlib is used to generate charts and graphs, such as user behavior heat maps, trend line graphs, etc.

[0075] Based on data analysis, the K-means clustering algorithm is used to group users. First, the basic attributes (including age, gender, region, etc.), behavioral characteristics (including browsing frequency, stay time, etc.) and preference characteristics (including service type preference, consultation topic, etc.) of users are extracted from the user behavior data to construct a feature vector. Then, the K-means algorithm is used to perform cluster analysis on the feature vector to group users with similar behavior patterns and interest preferences into one category. By continuously adjusting the number of cluster centers and the number of iterations, a reasonable user clustering model is finally obtained.

[0076] Build a detailed user portrait for each user group. Through in-depth analysis of the group's feature description, behavior patterns, and interest preferences, form a text description and extract key feature tags. These tags will help understand the characteristics and needs of each user group. In this embodiment, users are divided into different groups such as "high-frequency service users", "active consultation users", and "newly registered users", and corresponding service optimization strategies are formulated for each group.

[0077] Finally, a visual user portrait report will be generated, and the key information of the user portrait will be displayed in the form of charts, graphs, etc. Iterative optimization will also be carried out based on the accuracy and practicality feedback of the user portrait, and the construction method and model of the user portrait will be continuously improved to ensure that it always meets business needs and market changes.

[0078] In one embodiment, a deployment method for implementing multi-application integration based on the middle platform architecture is also included: First, we installed the system and configured the environment. We selected CentOS as the server operating system and installed it using the ISO image file. Then, we configured the JDK environment variables, installed the Tomcat container, Nacos service, and Redis service, and provided the necessary environment support for the application to run.

[0079] Next, we deployed the front-end and back-end services. For the front-end service, we installed Node.js and npm, and used Git to pull the front-end project code. We used the npm command to install project dependencies and packaged them to generate a dist package, and then copied the contents of this package to the webapp directory of Tomcat, so that the front-end resources can be accessed through the Web service. For the back-end service, we uploaded the compiled jar package to the server, started the service with the nohup and java -jar commands, and set the JVM memory parameters to ensure the stable operation of the application.

[0080] In order to improve the availability and response speed of the system, the Nginx server is configured as a reverse proxy, and a weighted polling algorithm is used to achieve intelligent load balancing of services. At the same time, a domain name certificate is configured, the domain name is pointed to the Nginx server, and the necessary ports are opened to ensure that external requests can be correctly distributed to the main server.

[0081] During the system data initialization phase, by accessing the management terminal homepage address and logging into the management system, basic data such as parameter information and dictionary information were configured, providing data support for the subsequent operation of the system.

[0082] In terms of security, a unified encryption strategy was implemented, sensitive data stored in the server cluster was encrypted, and the SSL / TLS protocol was configured to encrypt all data transmitted over the network. At the same time, a firewall was installed, an intrusion detection system was configured, and access control and authentication mechanisms were implemented, including setting up user accounts, configuring strong password policies, enabling multi-factor authentication, and setting up gateway tokens, which comprehensively improved the system's security protection capabilities.

[0083] Finally, in order to achieve interoperability between the VPC private network and the cloud layer, a cloud host instance was created in the VPC and the corresponding network settings were configured. The connectivity between the VPC and the cloud layer was tested using tools such as ping and traceroute, and the effectiveness of the routing strategy was verified. In addition, security testing tools were used to verify the effectiveness of the encryption strategy to ensure security during data transmission.

[0084] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A system for implementing multi-application integration based on a middle-end architecture, characterized in that: include: The middle office service module is used to build a middle office service layer that supports service calls and data sharing across multiple government applications; The resource management module uses virtualization technology to dynamically allocate hardware resources, encapsulate physical server resources into virtual machine instances, deploy cloud load balancers, and dynamically adjust resource allocation between virtual machine instances based on the real-time load of government applications; The security control module implements a unified user login password management strategy, formulates a user account lock strategy, sets a password validity period, and uses the AES symmetric encryption algorithm for data transmission encryption; The interface specification module adopts a unified docking method and formulates unified interface specifications, data exchange formats and business logic models; The data analysis module collects and analyzes user behavior data, embeds tracking code in the pages of government applications, records user behavior data, cleans the collected data, removes invalid and duplicate data, and mines and analyzes the cleaned data.

2. The system according to claim 1, characterized in that The framework of the middle platform service layer specifically includes: The front-end presentation layer uses a combination of Vue, Flutter and WeChat API frameworks to support multi-platform display; The backend logic layer is based on the SpringCloud microservice architecture and combines the RabbitMQ message queue to achieve efficient service calls; The persistence layer uses the cluster mode of PostGreSql database combined with Redis cache and ElasticSearch search engine.

3. The system according to claim 1, characterized in that The dynamic allocation of hardware resources is specifically implemented as follows: Build a virtualized resource pool to encapsulate physical server resources into independent virtual machine instances; Deploy a cloud load balancer to monitor the load of each virtual machine in real time and automatically adjust resource allocation using a weighted round-robin algorithm; Introduce a cache system to reduce direct database access; Set resource usage thresholds. When any resource metric exceeds or falls below the threshold, resource expansion or reduction operations are automatically triggered.

4. The system according to claim 3, characterized in that The automatic triggering of resource expansion or reduction operation is specifically as follows: When the resource load is detected to exceed the preset upper limit, a new virtual machine instance is automatically started to expand resources; When resource usage continues to be lower than the preset lower limit, idle virtual machines are automatically shut down and resources are released.

5. The system according to claim 1, wherein: The unified password management strategy is specifically implemented as follows: Integrated account lock function, setting lock conditions and automatic reset mechanism; Dynamically generate a unique AES encryption key for each integrated government application to ensure data transmission security.

6. The system according to claim 1, characterized in that The unified docking method includes interface docking, and the specific implementation steps include: Create government applications and automatically generate government application identifiers; Add government application version and submit for review; The administrator reviews the government application version. If it passes, the version is enabled and the government application configuration is performed. After the configuration is completed, it is displayed in the APP; if it fails, the current process is terminated and the government application is recreated; Create an interface and submit it for review; The administrator reviews the interface. If the review passes, the interface is bound to the enabled government application; if it fails, the current process ends and the interface is recreated.

7. The system according to claim 1, characterized in that The unified docking method also includes URL docking, and the specific implementation steps include: Create government applications; Add the link version and submit for review; The administrator reviews the link version. If it passes, the version is enabled and the government application is configured. After the configuration is completed, the link is displayed in the APP; if it fails, the current process is terminated and the government application and its link version are recreated.

8. The system of claim 1, wherein: The specific steps of collecting and analyzing user behavior data include: Collect user behavior data through embedding technology on government application pages and store the data in the database; Clean the data, remove noise, fill missing values ​​and perform standardization; Analyze data and build user portraits; Verify the accuracy of user portraits and perform iterative optimization based on feedback.

9. The system according to claim 8, characterized in that The step of constructing a user portrait further includes: Extract user basic attributes, behavior characteristics and preference characteristics, and construct feature vectors; Use clustering algorithms to group users and form a user grouping model; Build detailed user profiles for each user group, including feature descriptions, behavior patterns, and interest preferences; Generate a visual user portrait report.

Images