The Design and Implementation of Linux Containers Management System for Host Cluster

• Main Authors: Jou-yin Chen, 陳柔尹
• Other Authors: Chun-Hung Lin
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/rhw9r8

碩士 === 國立中山大學 === 資訊工程學系研究所 === 104 === In recent years, Big Data, distributed computing and parallel computing became popular. The traditional method of calculation has been unable to afford the huge amount of data computing. Through a combination of multiple servers into a cluster. Using cluster computing to improve performance and expand storage space. But it must afford the expensive servers, physical machine space, labor costs, low utilization of these costs do not meet the interests. So companies gradually replacing server via a small PC. Use virtualization technology to construct the distributed computing architecture. Virtualization technology can help developers quickly build independent of each other isolated containers. Container is using the system hardware and software resources, play host to the maximum efficiency. Through the powerful API and simple tool to easily create and manage the system, it has the advantage of efficiency and flexibility. Our system is designed to construct a decentralized architecture. Implement a system that provides the user to set up and manage a cluster architecture. Use containers to develop software, test environment or data analysis. The system uses LXC, and Container like a small machine. The user can enter the Container to test the environment, upload and download data or other application programs. We use Node.js construct a simple-to-use platform. Express framework to implement RESTful API, then execute Shell Script. To help users reduce the complexity of process architecture environment. As a result, users have a lower implementation costs, more flexible expansion capabilities, huge computing power, and improve the performance of the machine.