基於KubeEdge邊緣運算微服務部署自動化之系統設計

Designing a KubeEdge-based Edge Computing System with Microservices Deployment Automation

黃堤瑋(Ti-Wei Huang)

指導教授 ： 陳世穎

國立臺中科技大學/資訊與流通學院/資訊工程系碩士班/碩士(2020年)

摘要

隨著物聯網技術的發展，使得在網路邊緣端的連線設備數量快速增加，因此必須能即時有效地處理這些設備所產生的大量數據。傳統雲端架構會將資料傳至雲端的資料中心進行運算後，並將運算後的結果回傳至使用者設備，因此消耗大量頻寬，對網際網路與雲端基礎架構將造成極大的負荷。邊緣運算將原本在於雲端的運算擴展至邊緣縮短了與終端裝置的距離，因此得以降低網路的延遲和縮短了服務的回應時間，也減少頻寬的使用。本研究以邊緣運算概念將微服務部署至邊緣端，並採用開放原始碼系統建置一套基於KubeEdge邊緣運算微服務部署自動化系統，微服務所提出解決跨網域雲邊部署與維運的問題，KubeEdge以WebSocket的機制來串連雲端與邊緣端的連結，提供在不同的區域環境下雲端與邊緣端跨網域溝通方式，並提出微服務自動部署機制給KubeEdge容器對部署微服務的應用服務關係或管理。同時，為了確保系統的正常運作，本研究所提出的系統也提供雲端和邊緣端核心元件的資源監控。實驗部分測試部署邊緣端成功率，以驗證本系統部署為服務至邊緣端的可行性。並比較雲端與邊緣端之間的延遲對部署的影響。且本系統支援邊緣節點自我修復功能機制，為遇到節點重啟時能花費較短的時間恢復服務。

關鍵字

邊緣運算 ； 微服務 ； 容器

並列摘要

With the development of the Internet of Things technology, the number of connected devices at the edge of the network is rapidly increasing. Therefore, it is necessary to be able to effectively process the large amount of data generated by these devices in real time. The traditional cloud architecture transfers data to the cloud data center for computation, and returns the result to the end devices. By this way, the job consumes a large amount of bandwidth and will cause a heavy loading on the Internet and cloud infrastructure. Edge computing deploys the computing job that was originally in the cloud to the edge to shorten the distance from the end device, thereby reducing network latency and shortening the response time of the service, and also reducing the use of bandwidth. This study uses the concept of edge computing to deploy microservices to the edge, and uses an open source system to build an automated system for deployment of microservices based on KubeEdge edge computing to solve the problem of cross-domain cloud side deployment and maintenance. KubeEdge uses WebSocket mechanism is to connect the cloud and the edge to provide a cross-domain communication method between the cloud and the edge in different regional environments. The proposed system is used to deploy and manage microservices on the edge. Fuethermore, to ensure the normal operation of the system, the proposed system provides resource monitoring the core components between cloud and edge. The experimental part tests the success rate of deployment at the edge to verify the feasibility of deploying the microservices configured in the proposed system to the edge. And, the impact of the delay between the cloud and the edge on deployment is compared. In addition, the proposed system supports the self-healing function mechanism of edge nodes, which can take a short time to restore services when a node restarts.

並列關鍵字

Edge Computing ； Microservices ； Container ； KubeEdge

參考文獻

[1]M. Villari, M. Fazio, S. Dustdar, O. Rana and R. Ranjan, "Osmotic Computing: A New Paradigm for Edge/Cloud Integration", IEEE Cloud Computing, vol. 3, no. 6, pp. 76-83.

[2]P. Di Francesco, "Architecting Microservices", 2017 IEEE International Conference on Software Architecture Workshops (ICSAW).

[3]E. Kristiani, C. Yang, C. Huang, Y. Wang and P. Ko, "The Implementation of a Cloud-Edge Computing Architecture Using OpenStack and Kubernetes for Air Quality Monitoring Application", Mobile Networks and Applications.

[4]Z. Zhao et al., "A Novel Framework of Three-Hierarchical Offloading Optimization for MEC in Industrial IoT Networks", IEEE Transactions on Industrial Informatics, vol. 16, no. 8, pp. 5424-5434.

[5]Y. Xiong, Y. Sun, L. Xing and Y. Huang, "Extend Cloud to Edge with KubeEdge", 2018 IEEE/ACM Symposium on Edge Computing (SEC).