School of Electronics and Information Engineering, Tongji University, Shanghai 200092, China.
Sensors (Basel). 2023 Jun 26;23(13):5939. doi: 10.3390/s23135939.
In recent years, cloud-native technology has become popular among Internet companies. Microservice architecture solves the complexity problem for multiple service methods by decomposing a single application so that each service can be independently developed, independently deployed, and independently expanded. At the same time, domestic industrial Internet construction is still in its infancy, and small and medium-sized enterprises still face many problems in the process of digital transformation, such as difficult resource integration, complex control equipment workflow, slow development and deployment process, and shortage of operation and maintenance personnel. The existing traditional workflow architecture is mainly aimed at the cloud scenario, which consumes a lot of resources and cannot be used in resource-limited scenarios at the edge. Moreover, traditional workflow is not efficient enough to transfer data and often needs to rely on various storage mechanisms. In this article, a lightweight and efficient workflow architecture is proposed to optimize the defects of these traditional workflows by combining cloud-edge scene. By orchestrating a lightweight workflow engine with a Kubernetes Operator, the architecture can significantly reduce workflow execution time and unify data flow between cloud microservices and edge devices.
近年来,云原生技术在互联网公司中流行起来。微服务架构通过分解单个应用程序来解决多个服务方法的复杂性问题,从而使每个服务都可以独立开发、独立部署和独立扩展。同时,国内工业互联网建设仍处于起步阶段,中小企业在数字化转型过程中仍然面临着许多问题,如资源整合困难、控制设备工作流程复杂、开发和部署过程缓慢以及运维人员短缺。现有的传统工作流架构主要针对云场景,消耗大量资源,无法在边缘的资源受限场景中使用。此外,传统工作流传输数据的效率不够高,往往需要依赖各种存储机制。在本文中,通过结合云边场景,提出了一种轻量级、高效的工作流架构,通过编排一个带有 Kubernetes Operator 的轻量级工作流引擎,该架构可以显著减少工作流执行时间,并统一云微服务和边缘设备之间的数据流程。