XLCMS: a scalable and distributive Linux virtual cluster management system

  • Authors:
  • Tseng-Chang Yen;Felix Hsu;Shang-Juh Kao

  • Affiliations:
  • Department of Applied Mathematics, National Chung-Hsing University, Taichung, Taiwan;Department of Applied Mathematics, National Chung-Hsing University, Taichung, Taiwan;Department of Computer Science, National Chung-Hsing University, Taichung, Taiwan

  • Venue:
  • ICOIN'09 Proceedings of the 23rd international conference on Information Networking
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

As the computing environment moves toward integrating a group of computing powers, clusters have been increasing used by network services (for example, SourceForge.net and the UK JANET Web Caching Service). The ample resources of a modern computer make using virtualizations possible at all levels (system, storage, and network). By implementing at the user's level, User-Mode Linux (UML) allows users to run virtual machines as user processes. UML is primarily a commandline oriented virtual machine which provides several key/value pairs as arguments to the kernel to configure a virtual machine. Although useful for software developments, the complicated configuration set up of UML increases the user's work load. Several UML virtualization tools such as Netkit and VNUML have been proposed to simplify practicing UML. However, these tools usually sutTer from the ftexibility of use. Moreover, up and down of all virtual machines specified in the same configuration file must be booted or halted simultaneously which restricts the capability of separated management. In-depth knowledge of Linux system is actually required to effectively manage a Linux cluster. In this paper, we successfully develop a ftexible XLink-based Linux virtual Cluster Management System (XLCMS). Through the development of the XLCMS equipped with interactive, scalable and distributive management, we also employ XML-related techniques to process XLink configuration procedures for virtual machines. We are able to manage cluster composed of UML virtual machines from anywhere and dynamically execute tasks on individual virtual machines.