Design and Implementation of Highly Available Linux Clusters

نویسندگان

  • Ibrahim Haddad
  • Frederic Rossi
چکیده

Fueled by the Internet revolution and the lure of opportunities in the new converging world of voice, data and video over cost-effective networks, telecom and networking companies are attempting to offer the most complete and compelling end-to-end enterprise solutions. The interest in clustering from the telecom world comes from the fact that we can address the availability and scaled performance using cost-effective hardware and software while maintaining near telecom-grade characteristics. These characteristics include continuous service availability, high reliability, superior performance, high throughput for fast and reliable data streaming, flexibility in terms of fast reconfiguration, linear scalability, and ease and completeness of management. In 2000, the ARIES (Advanced Research on Internet E-Servers) project started at Ericsson Research Canada, aimed at finding and prototyping the necessary technology to prove the feasibility of a clustered Internet Server that provides telecom-grade characteristics. In 2001, ARIES evolved into a new direction to enhance the clustering capabilities Linux to fulfill the future demands for Mobile Internet Servers. This tutorial will address in detail all the design and implementation issues we faced building Linux clusters using Linux and Open Source Software as the base technology; in addition it will provide a how-to for building Linux clusters starting from scratch. Ibrahim Haddad is a System Designer at the Ericsson Research Open Architecture Lab in Montreal, Canada where he is primarily involved in researching carrier-class server nodes for real-time all-IP networks. The focus of his work at Ericsson is to bring telecom grade characteristics to Linux so that it can be considered as a potential operating system for carrier-class servers. Ibrahim received his Bachelor and Masters degrees in Computer Science from the Lebanese American University and he is currently a Dr.Sc. Candidate at Concordia University. Ibrahim is also involved in several Open Source projects and is a regular contributor to the Linux Journal. Frederic Rossi is a System Designer at the Ericsson Research Open Architecture Lab in Montreal, Canada where he is researching new load balancing and traffic distribution algorithms to improve Linux capabilities to be considered as a platform of choice for telecom grade clusters. Frederic received his Bachelor and Masters degrees in Computer Science from the University of Paris VIII. Proceedings of the 2001 IEEE International Conference on Cluster Computing (CLUSTER’01) 0-7695-1116-3/02 $17.00 © 2002 IEEE

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Linux Virtual Server for Scalable Network Services

This paper describes the motivation, design, internal implementation of Linux Virtual Server. The goal of Linux Virtual Server is to provide a basic framework to build highly scalable and highly available network services using a large cluster of commodity servers. The TCP/IP stack of Linux kernel is extended to support three IP load balancing techniques, which can make parallel services of dif...

متن کامل

On the Implementation and Performance of the (a, t) Protocol on Linux

This paper details the design and implementation of the ( ; t) protocol for ad-hoc networks on Linux. The ( ; t) protocol utilizes adaptive clustering to organize nodes into clusters in which the probability of path failure due to node movement can be bounded over time. Based on the ( ; t) cluster scheme, routes within clusters are maintained on a proactive basis while routing between clusters ...

متن کامل

Nswap: A Network Swapping Module for Linux Clusters

Transparency Users should not need to do anything special to use Nswap Non-invasive implementation How: implemented as a loadable kernel module Speed runs entirely at kernel level minimum overhead in time and space for protocols and state data Scalability Must scale well between small and large clusters How: Point-to-Point design: no central server Nodes make remote caching information using in...

متن کامل

The Design and Implementation of Berkeley Lab’s Linux Checkpoint/Restart

Clusters of commodity computers running Linux are becoming an increasingly popular platform for highperformance computing, as they provide the best price/performance ratio in the marketplace. But while the size and raw power of Linux clusters continues to increase, many aspects of their software environments continue to lag behind those provided by proprietary supercomputing systems. One featur...

متن کامل

Dodo: A User-level System for Exploiting Idle Memory in Workstation Clusters

In this paper, we present the design and implementation of Dodo, an e cient user-level system for harvesting idle memory in o -the-shelf clusters of workstations. Dodo enables data-intensive applications to use remote memory in a cluster as an intermediate cache between local memory and disk. It requires no modi cations to the operating system and/or processor rmware and is hence portable to mu...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2001