Search

CN-122018938-A - Cloud edge cooperation-oriented automatic software distribution and deployment method

CN122018938ACN 122018938 ACN122018938 ACN 122018938ACN-122018938-A

Abstract

The invention discloses a cloud-edge-collaboration-oriented automatic software distribution and deployment method and system, and belongs to the technical field of distributed computing and continuous software delivery. The cloud end establishes tasks based on declarative deployment description units and release strategies, an intelligent scheduler builds a hybrid distribution network to conduct differential scheduling according to node dynamic contexts, an edge agent instantiates the description units through environment injection, executes deployment through atomization transactions after pre-verification, decides flow switching or fast rollback according to multi-level health probe verification results, and the cloud end converges global states through bidirectional channels to achieve strategy-based closed-loop management and control. The cloud edge collaborative software delivery method solves the problems of low software distribution efficiency, complex deployment and adaptation, poor reliability and difficult collaborative management and control of mass heterogeneous edge nodes in a weak network environment, and achieves efficient, intelligent and highly reliable cloud edge collaborative software delivery.

Inventors

  • YU JIANXIN
  • WEI YONG
  • PEI XIAOQIANG
  • CUI JUNFENG
  • AN DAWEI

Assignees

  • 中国人民解放军63729部队

Dates

Publication Date
20260512
Application Date
20260128

Claims (9)

  1. 1. A cloud edge collaboration-oriented automatic software distribution and deployment method is characterized by running in a three-layer system architecture comprising a cloud management and control center, an edge service layer and an edge node agent, and comprising the following steps: The method comprises the following steps that S1, a cloud management and control center creates a declarative deployment description unit and a release strategy according to user input, and associates software products to generate a release task, wherein the deployment description unit defines a deployment blueprint by using structured data, and the deployment description unit comprises a target node constraint set, software product reference and selection logic, an atomization deployment execution plan, a multi-level health verification rule and an automation rollback plan; S2, an intelligent dispatcher in the cloud management and control center responds to the release task, builds a hybrid distribution network topology based on the acquired dynamic context information of the target edge node, and issues a differentiated distribution instruction to the target node; s3, after pulling and verifying the software product according to the received distribution instruction, the edge node agent performs environment injection and plan instantiation and analyzes the deployment description unit into an instantiation deployment instruction sequence aiming at the current node; s4, the edge node agent performs pre-deployment compliance verification in an isolation environment, executes the instantiation deployment instruction sequence in an atomization transaction mode after verification is passed, and performs health verification according to the multi-level health verification rule after execution is completed; s5, after the health verification is passed, performing production flow switching, cleaning temporary resources and reporting a successful state, and if the production flow switching fails at any stage of pre-verification, transaction execution or health verification, performing rollback operation according to the automatic rollback plan; s6, the cloud management and control center gathers the states of all edge nodes through a bidirectional state synchronization channel to form a global situation view, and performs state analysis and closed-loop management and control based on a preset strategy rule.
  2. 2. The cloud-edge collaboration-oriented automatic software distribution and deployment method of claim 1, wherein software product reference and selection logic in the deployment description unit supports dynamic selection of a best-matched product version according to tag attributes of target nodes, and the deployment stage defined by the atomized deployment execution plan comprises environment verification, resource preparation, product installation, configuration rendering, service start-stop and flow switching.
  3. 3. The cloud edge collaboration-oriented automatic software distribution and deployment method of claim 1 is characterized in that in step S2, the construction of the hybrid distribution network topology specifically comprises the steps of automatically organizing a point-to-point transmission network and designating seed nodes for clusters with good network quality and dense nodes, guiding to a nearest edge service layer cache or content distribution network node for scattered or cross-regional nodes, and generating an encrypted offline package and marking as a task to be synchronized for offline nodes.
  4. 4. The method for automatically distributing and deploying cloud-edge oriented collaboration software according to claim 1, wherein in step S4, the executing in an atomized transaction mode comprises creating checkpoints before and after key steps to ensure that directory switching operations are atomic operations and all intermediate operations are completed in a sandbox.
  5. 5. The method for automatically distributing and deploying cloud computing-oriented software according to claim 1, wherein the multi-level health verification rule comprises a start probe for judging whether a service process is successfully started, a ready probe for judging whether the service is ready to receive service traffic, and a service logic probe for verifying whether a core service function is normal.
  6. 6. The method for automatically distributing and deploying cloud-edge collaboration-oriented software according to claim 1, wherein in step S6, the closed-loop management and control comprises the steps that a policy orchestration engine automatically analyzes a state flow and triggers an automatic action when a preset condition is met, wherein the automatic action comprises pausing release, triggering rollback or generating an alarm.
  7. 7. A cloud-edge collaboration oriented automatic software distribution and deployment system for implementing the method of any of claims 1-6, comprising: The management and control center is deployed at the cloud and comprises a product warehouse for storing software products and a deployment description unit, a strategy arrangement engine for defining and analyzing a release strategy, an intelligent scheduler for planning a distribution path and a service for state monitoring and storage; the edge service layer is deployed in the regional data center and is used for providing software product caching, status reporting aggregation and limited local autonomous management; The edge node agent is deployed on each edge device and comprises a lightweight deployment engine for executing a deployment description unit and managing the life cycle of software; And the cloud management and control center, the edge service layer and the edge node proxy are communicated through a bidirectional state synchronization channel.
  8. 8. The method for automatically distributing and deploying cloud edge-oriented collaboration software of claim 7 wherein the edge node agent further comprises a download manager supporting continuous pulling of software products from multiple source break points according to priority, and integrity verification and signature verification.
  9. 9. The cloud edge collaboration-oriented automatic software distribution and deployment method of claim 7, wherein the cloud management and control center further comprises a unified management console for providing a global visual view, a publishing task board, node detail query and manual intervention interface.

Description

Cloud edge cooperation-oriented automatic software distribution and deployment method Technical Field The invention relates to the technical fields of distributed computing, edge computing and software continuous delivery, in particular to a method and a system for realizing automatic, intelligent and highly reliable distribution and deployment of software application, service components or artificial intelligent models to edge computing nodes with large scale, heterogeneous, limited resources and unstable network conditions in a computing architecture of cooperative work of a cloud end and an edge end. Background With the wide application of internet of things, industrial internet, 5G and artificial intelligence technologies, the computing paradigm is evolving from centralized cloud computing to "cloud-edge-end" collaborative computing. In the architecture, a cloud center is responsible for global data aggregation, complex model training and service arrangement, and an edge side is composed of massive computing nodes distributed in a near field of a physical world and is responsible for real-time data processing, low-delay response and local intelligent decision. The efficient and reliable synchronization of the software capability from the cloud to the edge side is a key for playing the cloud-edge cooperative value. However, the process faces serious challenges which are difficult to deal with by the traditional technology, namely, the contradiction between scale and efficiency, the requirement of concurrent deployment of millions of nodes far exceeds the manual or scripted operation and maintenance capability, the network environment is complex, the edge nodes are always positioned in wide area networks with limited bandwidth and unstable connection, the transmission of large-volume software packages is extremely easy to fail, the third, the nodes are highly heterogeneous, the software adaptation and deployment are extremely complicated due to the differences of hardware architecture, operating system and resource specification, the fourth, the reliability guarantee is weak, the edge side environment is complex, the risk of deployment failure is high, an effective automatic recovery mechanism is lacked, and the fifth, the cooperative management and control is difficult, so that the existing tool lacks the capability of carrying out fine gray scale release, state synchronization and closed-loop control on the large-scale edge nodes. At present, although a device management scheme based on container arrangement exists, the design is mainly oriented to isomorphic and high-bandwidth data center environments, and when dealing with large-scale, weak connection and high-isomerism characteristics of edge scenes, the device management scheme has obvious defects in distribution efficiency, environment adaptation, deployment reliability and collaborative management and control. Therefore, a method for automatically distributing and deploying systematic software specially designed for cloud-edge collaborative design is needed. Disclosure of Invention The invention aims to overcome the defects of the prior art and provides a cloud-edge-oriented collaborative software automatic distribution and deployment method and system. The method aims at achieving the core aims of achieving efficient and intelligent distribution of software products to a super-large-scale edge node cluster, completing automatic and seamless deployment and adaptation of software in a heterogeneous edge environment, constructing a complete mechanism for guaranteeing high reliability and service continuity in a deployment process, and establishing a cloud edge cooperative management and control system supporting fine operation and maintenance. In order to achieve the aim of the invention, the invention provides a cloud edge collaboration-oriented automatic software distribution and deployment method. The method is characterized in that an automatic, observable and controllable software delivery closed loop from the cloud end to the edge end is constructed through a declarative deployment description unit, an intelligent coordination engine and a bidirectional state synchronization channel. 1. System architecture and core component The method operates in a three-tier system architecture comprising a cloud management and control center, an edge service layer, and an edge node proxy. The cloud management and control center comprises a product warehouse, a strategy arrangement engine, an intelligent dispatcher and a state monitoring service, wherein the edge service layer is used as a regional relay to provide buffering and aggregation functions, and the edge node agent resides on each edge device and comprises a lightweight deployment engine. The components communicate and are state synchronized via a bi-directional state synchronization channel. 2. Deployment intent abstraction based on declarative descriptions The present invention introduces a machi