Building a Large Scalable Internet Superserver for Academic Services with Linux Cluster Technology

نویسندگان

  • Putchong Uthayopas
  • Surasak Sanguanpong
  • Yuen Poovarawan
چکیده

With the speed and bandwidth offered by the next generation Internet technology, there is a need for large and scalable Internet server that can provides an adequate computing power and storage for the new generation Internet applications. This requires a huge investment in a very large and expensive commercial server system. Recently, the emergence of Linux PC clustering or so-called Beowulf Cluster Computing technology has provided a very low cost, scalable, and high-performance alternative to using these commercial superserver system in many class of applications. This paper presents our experiences learned from the construction of a 72 nodes Beowulf Class Cluster called PIRUN (Pile of Inexpensive and Redundant Universal Nodes) Cluster at Kasetsart University, Bangkok, Thailand. The purpose of this project is to explore the use Linux PC clustering technology to build a central Internet superserver and supercomputing class system for academics support at the university level. In this paper, the planning of the system is presented along with the discussion about issues in both hardware and software. This system when fully exploited will cost the same as current same main server but offer about 30-40 times more computing power and storage. Moreover, PIRUN System will becomes the largest supercomputing system in Thailand and deliver about 3 times higher performance than the previous largest supercomputing system in Thailand. We strongly believe that this type of system and technology is very suitable for developing countries in Asia Pacific Region.

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

Performance Analysis of a Cluster File System

Design of cluster file system is very important for building a general-purpose cluster with commodity components. To provide scalable high I/O performance needed in the scientific computing, engineering computing, Internet/Intranet services, I/O intensive commercial database and multimedia application, we designed a high-performance cluster file system named COSMOS. This paper presents the basi...

متن کامل

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

Clustering Support and Replication Management for Scalable Network Services

The ubiquity of the Internet and various intranets has brought about widespread availability of online services and applications accessible through the network. Cluster-based network services have been rapidly emerging due to their costeffectiveness in achieving high availability and incremental scalability. This paper presents the design and implementation of the Neptune middleware system that...

متن کامل

A 64-bit, Scalable File System for Storage Area Networks

By merging network and channel interfaces, resulting interfaces allow multiple computers to physically share the storage devices in storage area network (SAN). In SAN, computers service local file requests directly from shared storage devices. Direct device access eliminates the server machines as bottlenecks to performance and availability. Communication is unnecessary between computers, since...

متن کامل

TODS: Object Storage Platform Designed for Scalable Services

TODS is an object storage platform we are building to offers scalability, availability, performance and ease of use, to simplify the building of scalable Internet services. It enables the convergence of programming and storage model to Object-Orientation and provides transparent persistence for data objects, which greatly simplifies application development. Decentralized cluster design and effi...

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 2000