CN109857514B - Cross-platform, cross-commercial autonomous environment, complex and giant information system hybrid deployment system - Google Patents

Abstract

The invention relates to the field of computer software, and aims at a complex giant system composed of numerous functions/services/subsystems with complex installation and runtime invocation dependencies, and proposes a deployment system, which can be deployed across physical servers/virtual machines/containers and other different basic operating platforms, across commercial hardware environment based on X86 instruction set / Feiteng independent hardware environment based on ARM instruction set / Loongson independent hardware environment based on RISC instruction set / Shenwei independent hardware environment based on independent instruction set, cross windows/ In the deployment environment of Linux/Galaxy Kylin and the domestic operating system that won the bid for Kylin, there are web applications/RPC, gRPC, Webservice, Restful type services/Java applications/C applications/C++ applications/Firefox browser plug-ins/Dameng, Jincang Unified management of the deployment process of complex giant information systems in the form of domestic database data files/container images/virtual machine image software.

Description

Cross-platform, cross-commercial autonomous environment, complex and giant information system hybrid deployment system

technical field

The invention belongs to the field of computer software, and in particular relates to a hybrid deployment system of a complex giant information system that is cross-platform and cross-commercial autonomous environment.

Background technique

Complex giant information system, especially refers to complex composition (generally including hundreds of subsystems/services/functional domains) and complex deployment associations (there are one-to-many, many-to-many, one-to-many, many-to-many, and Strongly ordered local installation dependencies, and multi-level call chain relationships during runtime), huge size (over 10G in total in the system), complex deployment environment (cross physical servers/virtual machines/containers and other basic operating platforms, cross Commercial hardware environment based on X86 instruction set/Feiteng independent hardware environment based on ARM instruction set/Godson independent hardware environment based on RISC instruction set/Shenwei independent hardware environment based on independent instruction set, cross windows/Linux/Galaxy Kirin, won the bid for Kirin Domestic operating system), complex composition (the software form of each subsystem/service covers Web application, RPC/Grpc/Webservice/Restful type service, Java application, C application, C++ application, Firefox browser plug-in, Dameng/Jincang Domestic database data files, container images, virtual machine images). With the steady advancement of my country's core technology independent and controllable strategy, complex giant information systems are currently widely used in the military, scientific research and other fields to handle complex business (such as command and control).

However, when a complex giant information system is deployed and installed, it often faces a variety of internal components, which are difficult to deploy uniformly, the deployment environment is complex and difficult to plan and adjust, the internal dependencies of the system are cumbersome and difficult to arrange solutions, the large-scale concurrent deployment process is difficult to schedule efficiently, and the status of the installation and deployment process is difficult. Synchronous acquisition and other issues.

From the perspective of previous published papers, there are deployment solutions based on multiple server architectures across X86/SPARC ("Automated Deployment Server System" Network Security and Informatization 2018, (06), 98-100), and some are based on complex environments in specific fields. The proposed deployment scheme ("Commercial Bank Multi-system Complex Environment Version Release Automation Practice" China Financial Computer 2018, (04), 51-54), "The Practice of Automated Batch Deployment Management" (China Media Technology 2017, (10), 119 -121), there is an automated deployment scheme based on the cloud platform environment ("Design and Implementation of Cloud Platform Automated Deployment" Electronic Design Engineering 2018, 26(03), 60-64+68), there is an automated deployment scheme based on the Docker container environment ( "Research on Application Deployment Management Platform Based on Docker" Electronic Design Engineering 2017, 25(12), 41-44-48), there is an automatic deployment scheme for virtualized environment ("Application of Virtualization Automatic Deployment Key Technology" Experimental Technology and Management 2016, 33(08), 126-128+136), there is still a certain gap in the deployment management of complex and huge information systems, which is manifested in: first, most solutions focus on a specific problem domain or a specific system, For example, "The Practice of Automated Batch Deployment Management" (China Media Technology 2017, (10), 119-121) focuses on Xinhua News Agency's "Media Convergence Development Project"; second, the coverage of the basic operating platform in the deployment environment is incomplete, only It can be oriented to one or several types of basic operating platforms. For example, "Docker-based Application Deployment Management Platform Research" is only for container environments, and "Cloud Platform Automated Deployment Design and Implementation" is only for cloud platform virtual machine environments; the third is the deployment of The coverage of software forms is incomplete, and it is impossible to package and implement the installation and deployment process for different forms of software according to unified standards; Fourth, there is a general lack of support for independent hardware environments, and it is impossible to deploy and install on the current mainstream independent hardware platforms based on Feiteng 1500A; It is impossible to manage the installation and deployment process in a unified and simultaneous manner on the mainstream Windows and Linux operating systems, let alone support the installation and deployment on domestic independent operating systems such as Galaxy Kirin.

SUMMARY OF THE INVENTION

Aiming at the deficiencies of the prior art, the present invention provides a hybrid deployment system of complex and giant information systems across platforms and commercial autonomous environments, which can support different basic operating platforms such as physical servers/virtual machines/containers, and commercial applications based on X86 instruction sets. Hardware environment / Feiteng independent hardware environment based on ARM instruction set / Loongson independent hardware environment based on RISC instruction set / Shenwei independent hardware environment based on independent instruction set, cross-windows/Linux/Galaxy Kirin, won the bid for Kirin domestic operating system, inclusive Web Application, RPC/Grpc/Webservice/Restful type service, Java application, C application, C++ application, Firefox browser plug-in, Dameng/Jincang domestic database data file, container image, virtual machine image software form, which can be run on the basis of Unified management of complex and giant information systems deployed in the four-dimensional environment of platform, hardware environment, operating system, and software form.

In order to solve the above-mentioned technical problems, the present invention realizes through the following technical solutions:

A hybrid deployment system for complex giant information systems across platforms and commercial autonomous environments is characterized in that it includes a software package verification tool, a software repository, a deployment console, a deployment task scheduling, a deployment interface, a unified message framework, a deployment agent, and a registration center. The software package verification tool will run in different basic operating environments, different operating systems, and different commercial autonomous environments. Dream/Jincang domestic database data files, container images, and virtual machine images are packaged in accordance with unified packaging standards, and a packaging description file is configured to describe the basic requirements of the deployment environment, the basic requirements of the software operating environment, and other applications or services during deployment. Dependency relationship, calling relationship with other services during operation, the final output is a deployment package that can perform unified installation and deployment and upload it to the software warehouse; the software warehouse is the place where the deployment package is placed, and all packaged deployment packages After going through the process of package integrity verification, deployment description correctness verification, cataloging and classification, and manual review, the software repository can be adapted to complex deployment scenarios and supports hierarchical deployment, regular backup, discrete storage, and multiple versions. Management, pre-upload review, cataloging settings, label settings, and externally provide conditional query functions; the deployment console is where deployment strategies are set and deployment plans are generated. For all deployment nodes and related information, the user can establish a mapping relationship between the deployment package in the software warehouse and the deployment node according to actual needs. During the process of building the mapping relationship, the dependent applications or services will be automatically associated. , receive the deployment status feedback pushed by the deployment task scheduling and display it uniformly; the registration center perceives the active status of all the deployment agents, receives the registrations of the deployment agents of each version, forms a deployment node list, and periodically communicates with the deployment agents. The agent performs heartbeat connection for liveness detection, and the registration center can know the current running status of each deployment agent. Once the deployment agent exits abnormally or is damaged for other reasons, the registration center can immediately perceive the deployment The agent goes offline and becomes inactive, and timely informs the deployment task scheduling; the deployment task scheduling is the control center of the entire deployment system, responsible for parsing the mapping relationship between the deployment package and the deployment node set by the user in the deployment console, and Generates the corresponding deployment and installation instruction set, and is responsible for distributing deployment instructions to the deployment agents on all deployment nodes through the unified messaging framework at the beginning of deployment, to ensure that the local dependencies and network access dependencies of each application or service during the deployment and installation process are not. Correctly established, in addition, the current deployment agent activity status is obtained from the registration center, and the current deployment process status is obtained from the deployment agent. Once a deployment agent is found to be offline, an alarm prompt will be issued immediately, and the deployment will be temporarily suspended The installation process will not occur until the deployment agent is brought back to an active state Continue to perform deployment and installation; the deployment interface provides query services for the deployment agent during the deployment process, provides basic information of the deployment package package, local installation, and network operation access dependency information obtained from the software warehouse, and also receives the deployment process. related information, and transmit the received information to the deployment task scheduler; the unified messaging framework provides a message sending and receiving mechanism for each component of the deployment system, and provides the deployment console, the deployment task scheduler, the deployment interface , The deployment agent provides a message sending and receiving channel, and the unified message framework provides two message sending and receiving mechanisms, synchronous and asynchronous, as well as a retransmission mechanism after sending failure; the deployment agent is specifically responsible for the deployment and installation work, which is divided into multiple version, which can run on different basic operating platforms, different hardware environments, and different operating systems. The deployment agent automatically connects to the registration center, receives the heartbeat detection initiated by the registration center, and responds in time. The unified messaging framework receives the instructions sent by the deployment task scheduling, unpacks and executes them, and calls the deployment interface during the execution process to obtain the basic information, local installation, and network operation access dependencies of the encapsulated package. During the process, the current deployment situation is fed back to the deployment task scheduling through the unified messaging framework.

In the above technical solution, preferably, the software encapsulation verification tool can encapsulate and describe applications or services of different technical forms according to a unified standard.

In the above technical solution, preferably, the software warehouse can be deployed at multiple locations in a hierarchical manner, and various types of data stored in the software warehouse are discretely stored.

In the above technical solution, preferably, the task scheduling can ensure the correct establishment of the local dependencies and network access dependencies of each application or service during the deployment and installation process.

In the above technical solution, preferably, the task scheduler will give an alarm prompt once it finds that a certain deployment agent is offline during the deployment and installation process, and temporarily suspend the deployment and installation process until the deployment agent is restored to the active state. The deployment installation will continue.

In the above technical solution, preferably, the deployment agent can be divided into C++ or Java version from the development language, and can be divided into independent environment adaptation or commercial environment adaptation version from the technical system, which can be compatible with independent or commercial environment during deployment. Environment, virtual machine or container and various basic operating platforms, commercial or domestic operating systems.

In the above technical solution, preferably, the deployment agent is automatically registered with the registration center after being started.

In the above technical solution, preferably, the deployment interface receives the relevant information of the deployment process including a log of the current client application or service installation progress, installation location, whether the installation is completed, and whether the installation process is wrong.

In the above technical solution, preferably, different basic operating platforms in the deployment agent include physical machines, commercial cloud platforms, commercial container clouds, autonomous cloud platforms, and autonomous container clouds; different hardware environments include autonomous Feiteng series, autonomous Loongson series , Independent Shenwei series, commercial X86 series; different operating systems include Windows series, Linux series, domestic Galaxy Kirin, domestic winning Kirin.

The present invention proposes a deployment system for a complex giant system composed of numerous functions/services/subsystems with complex installation and runtime calling dependencies, which can support different basic operating platforms such as physical servers/virtual machines/containers. , Cross-commercial hardware environment based on X86 instruction set / Feiteng independent hardware environment based on ARM instruction set / Loongson independent hardware environment based on RISC instruction set / Shenwei independent hardware environment based on independent instruction set, cross windows/Linux/Galaxy Kirin, Winning the bid for Kylin domestic operating system, inclusive web application, RPC/Grpc/Webservice/Restful type service, Java application, C application, C++ application, Firefox browser plug-in, Dameng/Jincang domestic database data file, container image, virtual machine image The software form can realize the unified management of the deployment of complex and huge information systems in the four-dimensional environment of the basic operating platform, hardware environment, operating system, and software form.

The present invention is oriented to the deployment process of a giant information system in a complex environment, and the beneficial effects embodied are as follows:

a) Support deployment and installation across commercial and autonomous hardware platforms;

b) support deployment and installation across commercial and autonomous operating systems;

c) Support the deployment and installation of application software in various technical forms;

d) Support the deployment and installation of multiple operating support environments across physical servers/cloud platforms/containers;

e) Support the setting and verification of the two deployment dependencies of local installation dependencies and runtime invocation dependencies;

f) Supports automatic arrangement of deployment instruction sets according to the mapping of deployment nodes and applications to be deployed;

g) Support the unified display of the progress of the deployment process and the alarm prompt of abnormal errors;

h) Supports the export and import of deployment schemes, which facilitates rapid replication of off-site deployments.

Description of drawings

FIG. 1 is a design diagram of a deployment process model of the present invention.

FIG. 2 is a functional structure diagram of the deployment system of the present invention.

FIG. 3 is a logical diagram of the internal relationship and invocation of the deployment system of the present invention.

FIG. 4 is a flow chart of the implementation of the deployment system of the present invention.

Detailed ways

The technical solution of the present invention will be described in further detail below from the two aspects of model design and composition structure with reference to the accompanying drawings.

The model design of the present invention is shown in Figure 1, including software application packaging model, packaging integrity verification, deployment description correctness verification, software application warehouse model, executable deployment sequence, deployment environment conditions, application software dependency constraints to be installed, remote Automatic deployment and installation of eight parts.

The software application packaging model is used to describe the basic requirements of the software deployment environment (for example: running on the operating system/version, running on the X86/Feiteng 1500A architecture hardware platform), software itself information (for example: development language, version, research unit), software composition Information (for example: what program files, configuration files, directories are there), basic requirements of the software operating environment (for example: whether the Java operating environment/version is required, whether the C++ operating environment/version is required), the client application or service deployment and installation dependencies Other application and service information (for example: identification, name, version, commercial/autonomous), other service information to be called during service operation (for example: identification, name, version, commercial/autonomous).

Encapsulation integrity verification is to verify the encapsulated client or service based on the software encapsulation model to determine whether the content is complete; the deployment description correctness verification is based on the encapsulation model, to the deployment environment required by the encapsulated client or service. Verify that the content is complete and compliant.

The software application warehouse model is the definition of the encapsulated application or service storage structure, which defines distributed storage (storage files in discrete form), hierarchical storage (similar to the concepts of general library and sub-library, which can be based on some of the encapsulated files. The attribute specifies the storage location), regular backup (the warehouse is divided into the main site and the backup site, and synchronous backup is performed regularly and regularly), classification rules (for the package files stored in the warehouse, they can be classified and distinguished by labels and catalogs. , easy to give business meaning).

The deployment environment describes the specific deployment scenarios, such as: the number of deployment nodes, the number of X86 commercial machines, the number of Feiteng 1500A self-hosts, the number of Loongson 3A self-hosts, the number of Shenwei self-hosts, what operating system is installed on each deployment node, The computing/storage/network configuration of each deployment node.

The executable deployment sequence is a deployment execution plan generated based on actual deployment requirements and points in the current deployment environment, indicating which client applications or services are installed on which deployment points, what the installation sequence is, and what needs to be executed after the installation is complete. What initialization operation.

The software dependency constraint of the application to be installed is to re-check the installation dependencies of the client application or service to be installed in the current deployment node before deployment, so as to determine that the content to be installed on the local machine is all ready.

Remote automatic deployment and installation is a process in which the deployment tool on the deployment node receives the executable deployment sequence, parses the instructions, and executes the specific deployment process. Hardware environment / Feiteng independent hardware environment based on ARM instruction set / Loongson independent hardware environment based on RISC instruction set / Shenwei independent hardware environment based on independent instruction set, deploy across windows/Linux/Galaxy Kirin, and won the bid for Kirin's domestic operating system.

(2) Composition structure

The system structure of this patent is shown in Figure 2, which specifically includes eight parts: software packaging verification tool, software warehouse, deployment console, deployment task scheduling, unified message framework, registration center, deployment interface, and deployment agent.

The software package verification tool will run in different basic operating environments, different operating systems, and different commercial autonomous environments. Dream/Jincang domestic database data files, container images, and virtual machine images are packaged in accordance with unified packaging standards, and a packaging description file is configured to describe the basic requirements of the deployment environment, the basic requirements of the software operating environment, and other applications or services during deployment. Dependencies, calling relationships with other services during runtime, etc., and the final output is a deployment package that can perform unified installation and deployment.

The software warehouse is the place where the deployment packages are placed. All packaged deployment packages are finally put into the warehouse after the process of package integrity check, deployment description correctness check, cataloging and classification, and manual review. The software warehouse can adapt to complex Deployment scenarios, support hierarchical deployment, regular backup, discrete storage, multi-version management, and provide external conditional query functions.

The registration center perceives the active status of all deployment agents, which is achieved through the automatic registration of the deployment agents pre-placed on the deployment nodes. When the deployment agents are started, they will automatically register with the registration center and make regular heartbeat connections, so that the registration center can You can know the current running status of each deployment node. Once the deployment node is shut down or damaged for other reasons, the registry can immediately sense the offline status and notify the deployment task scheduling in time.

The deployment console is the place to set the deployment strategy and generate the deployment scheme. The console lists all the current deployment nodes and related information (ip, open ports, etc.) through perception and interaction with the deployment nodes. A mapping relationship is established between the deployment package and the deployment node, and the dependent applications or services are automatically associated during the process of building the mapping relationship; during the deployment process, the deployment status feedback pushed by the deployment task scheduling is received and displayed.

Deployment task scheduling is the control center of the entire deployment system. It is responsible for parsing the mapping relationship between the deployment package and deployment nodes set by the user in the deployment console, and generates the corresponding deployment and installation instruction set. The deployment agent distributes deployment instructions to ensure the correct establishment of local dependencies and network access dependencies of each application or service during the deployment and installation process, and also obtains the current deployment agent activity status from the registration center, and obtains from the deployment agent In the current state of the deployment process, once a deployment agent is found to be offline, an alarm prompt will be issued immediately, and the deployment and installation process will be temporarily suspended until the deployment agent is restored to the active state, and the deployment and installation will not continue.

The unified messaging framework provides a message sending and receiving mechanism for each component of the deployment system, and provides a message sending and receiving channel for the deployment console, deployment task scheduling, deployment interface, and deployment agents to achieve functional decoupling. At the same time, the unified messaging framework provides synchronous and asynchronous. Two message sending and receiving mechanisms, as well as a retransmission mechanism after sending failure, ensure high reliability of message sending and receiving.

The deployment interface provides query services for the deployment agent during the deployment process, and provides basic information about the deployment package, local installation, and network operation access dependencies. It also receives information related to the deployment process, such as the current client application or service installation progress, The installation location, whether the installation is complete, the error log of the installation process, etc., and transmit the received information to the deployment task scheduler.

The deployment agent is specifically responsible for the deployment and installation work. It is divided into multiple versions, which can run on different basic operating platforms such as physical servers/virtual machines/containers, and run on the commercial hardware environment of the X86 instruction set/the ARM-based Feiteng independent hardware environment. 、Running on different operating systems such as windows/Linux/domestic operating systems, the deployment agent receives the instructions sent by the deployment task scheduling through the unified messaging framework, unpacks and executes them, and calls the deployment interface during the execution process to obtain the basic information of the package. Information, local installation, and network operation access dependencies, and during the deployment process, the current deployment situation is fed back to the deployment task scheduling through the unified messaging framework.

(3) Implementation process

The realization flow of this patent is shown in Figure 4. The entire system deployment and installation process is as follows:

a) In the software development stage, developers will package and package the client applications or services that have passed the test and verification according to the unified standard to form a deployment package. After review, submit the software repository. Before submission, you can set labels, catalogs, classifications, and special topics. etc. to facilitate retrieval and query;

b) In the deployment preparation stage, security personnel first push the deployment agent to all deployment nodes through the deployment console (environments include physical server virtual machines/containers, commercial/independent hardware environments, Windows/Linux/domestic Galaxy Kylin/domestic Win the bid for Kylin), and execute the automatic installation script to execute the installation process and start;

c) After the deployment agent is started, it automatically registers with the registration center, and provides the IP address of its own deployment node, port opening, etc.; after the registration center receives the registration information of the deployment agent, it will periodically perform heartbeat detection to determine the activity of the deployment agent;

d) The deployment console obtains the active status of all current deployment agents through deployment task scheduling. When the deployment agent statuses of the deployment nodes that map the deployment installation package are all active, the premise that the current deployment environment is complete is considered to be established; by calling the software repository interface to view Whether all deployment installation packages and their local installation dependencies and runtime access dependencies exist, if they exist, the premise that the current deployment installation package is complete is considered to be true; if one of the above preconditions is not true, an alarm will be prompted;

e) If the above preconditions are all established, all current deployment nodes, as well as packaged client applications and services are listed in the deployment console, and the user can set the mapping between the two;

f) During the deployment work, the user initiates the deployment task at the deployment control end. At this time, the deployment task scheduling verifies the deployment environment again. When the deployment agent on the deployment node remains active, all deployment installation packages and their local dependencies When both conditions are met, the deployment task scheduling obtains the mapping relationship set by the user, and after analysis, a multi-channel executable deployment and installation instruction set sequence is formed;

g) Deployment task scheduling distributes the generated deployment instructions to the deployment nodes through the unified messaging framework. After the deployment agent receives the instructions, it first connects to the software repository to download the deployment package, and verifies the integrity of the package in turn according to the installation sequence set in the instruction. , Whether the local dependencies are ready for download and installation, if all conditions are met, start the deployment and installation, and during the installation process, the deployment status and error information of each application are fed back to the deployment task feedback through the unified messaging framework;

h) After the deployment task feedback summarizes the deployment status of each deployment node, it is pushed to the deployment control terminal to display the deployment progress uniformly; at the same time, if the local installation fails or the invocation of the dependent installation fails during the luck period, the security personnel will be alerted in time, and Temporarily suspends the deployment installation process until the deployment agent returns to an active state.

The protection scope of the present invention includes, but is not limited to, the above embodiments, and any substitutions, deformations, and improvements made to the present technology that are easily conceived by those skilled in the art all fall into the protection scope of the present invention.

Claims (4)

1. A hybrid deployment system for complex giant information systems across platforms and commercial autonomous environments, characterized in that it includes software encapsulation verification tools, software warehouses, deployment consoles, deployment task scheduling, deployment interfaces, unified messaging frameworks, deployment agents, and registration centers, The software package verification tool will run in different basic operating environments, different operating systems, and different commercial autonomous environments for Web applications, RPC/Grpc/Webservice/Restful type services, Java applications, C applications, C++ applications, and Firefox browser plug-ins , Dameng/Jincang domestic database data files, container images, and virtual machine images are packaged in accordance with unified packaging standards, and a packaging description file is configured to describe the basic requirements of the deployment environment, the basic requirements of the software operating environment, and other applications during deployment. Or the dependency relationship of the service, the calling relationship with other services during operation, and the final output is the deployment package that can perform unified installation and deployment and upload it to the software warehouse; the software warehouse is the place where the deployment package is placed, and the deployment of all packages The package package is finally put into the warehouse after the process of package integrity check, deployment description correctness check, cataloging and classification, and manual review. The software warehouse can adapt to complex deployment scenarios and supports hierarchical deployment, regular backup, discrete storage, Multi-version management, pre-upload review, catalog setting, label setting, and externally provide conditional query functions; the deployment console is where deployment strategies are set and deployment plans are generated. All the current deployment nodes and related information are displayed, and the user can establish a mapping relationship between the deployment package in the software warehouse and the deployment node according to actual needs. During the deployment process, the deployment status feedback pushed by the deployment task scheduling is received and displayed uniformly; the registration center perceives the active status of all the deployment agents, receives the registrations of the deployment agents of each version, forms a deployment node list, and regularly communicates with all deployment agents. The deployment agent performs heartbeat connection for liveness detection, and the registration center can know the current running status of each deployment agent. Once the deployment agent exits abnormally or is damaged for other reasons, the registration center can immediately perceive all the deployment agents. The deployment agent goes offline and becomes inactive, and timely informs the deployment task scheduling; the deployment task scheduling is the control center of the entire deployment system, responsible for parsing the mapping relationship between the deployment package and the deployment node set by the user in the deployment console , and generate the corresponding deployment and installation instruction set. When starting the deployment, it is responsible for distributing deployment instructions to the deployment agents on all deployment nodes through the unified messaging framework to ensure the local dependencies and network access dependencies of each application or service during the deployment and installation process. The correct establishment of the relationship, in addition to obtaining the current deployment agent activity status from the registration center, and obtaining the current deployment process status from the deployment agent, once a deployment agent is found offline, it will immediately give an alarm and prompt, and temporarily hang up Start the deployment and installation process until after the deployment agent is back to active state The deployment and installation will continue to be executed; the deployment interface provides query services for the deployment agent during the deployment process, provides the basic information of the deployment package package, local installation, and network operation access dependency information obtained from the software warehouse, and also receives deployment process related information, and transmit the received information to the deployment task scheduler; the unified messaging framework provides a message sending and receiving mechanism for each component of the deployment system, and provides the deployment console, the deployment task scheduler, the A message sending and receiving channel is provided between the deployment interface and the deployment agent, and the unified messaging framework provides two message sending and receiving mechanisms, synchronous and asynchronous, and a retransmission mechanism after sending fails; the deployment agent is specifically responsible for the deployment and installation work, and is divided into two types: Multiple versions can run on different basic operating platforms, different hardware environments, and different operating systems, and the deployment agent automatically connects to the registration center, receives the heartbeat detection initiated by the registration center, and responds in time, The instructions sent by the deployment task scheduling are received through the unified messaging framework, unpacked and executed, and the deployment interface is called during the execution process to obtain the basic information, local installation, and network operation access dependencies of the encapsulated package. During the deployment process, the current deployment situation is fed back to the deployment task scheduling through the unified messaging framework. 2. the hybrid deployment system of cross-platform, cross-commercial autonomous environment complex giant information system according to claim 1, is characterized in that, described deployment agent can be divided into C++ or Java version from development language, can be divided into from technical system as The self-adapted or commercial environment-adapted version is compatible with self-contained or commercial environments, virtual machines or containers, and various types of basic operating platforms, commercial or domestic operating systems during deployment. 3. The cross-platform cross-commercial autonomous environment complex giant information system hybrid deployment system according to claim 1, wherein the deployment interface receives the relevant information of the deployment process including the installation progress, installation location of the current client application or service , whether the installation is complete, and whether the installation process is wrong. 4. The hybrid deployment system for complex giant information systems in a cross-platform, cross-commercial autonomous environment according to claim 1, wherein the different basic operating platforms in the deployment agent include physical machines, commercial cloud platforms, commercial container clouds, autonomous Cloud platform, independent container cloud; different hardware environments include independent Feiteng series, independent Godson series, independent Shenwei series, commercial X86 series; different operating systems include Windows series, Linux series, domestic Galaxy Kirin, domestic winning Kirin.

Images