Rumon is named after Rumi. Who was Rumi?

    RUMON is a java-based complete solution for high performance computing on workstation clusters. These clusters are used for high-performance Beowulf class quassi-supercomputers. We are using such a cluster for scientific computation with parallel utilities for various physics and engineering problems. The RUMON is a tool that comprises a middleware that allows effective monitoring and control of system resources as well as interactivity and performance measurement for parallel programs. The end-user can submit an application and steer it remotely through the applet window. The tool has been tried and tested on 4-node Beowulf Cluster RUMI developed by the Khwarzimic Science Society cluster team. Built on top of CUMULVS and Pardyn (existing libraries for performance visualization) and completely developed in java, the tool fits for dedicated Linux clusters, Message passing libraries and wide range of classes of parallel applications.    

	Rumon will enable clients view system resources and manage the cluster remotely as well as interact with the submitted application.

Please visit Rajkumar Buyya's Cluster Computing Infoware. Most of the presentations and resources available on this compendium are also available in the KSS Office.

     All the reports regarding the project are available here.

    As the Performance of commodity computers increase and the network hardware prices reduce, it is much more feasible to run your resource intensive programs and execute high order calculations using parallel computational systems rather than buying time on high priced supercomputers. Beowulf class clustering is high prized too! This technology uses cheap microprocessors that are commercially and conveniently available and the price per performance ratio of a typical Beowulf machine is three to ten times better than that from traditional supercomputers.

  ds

Beowulf is a multi-computer architecture which can be used for parallel computations. It is a system which usually consists of one fileserver (at least) and one or more client nodes connected together via Ethernet or some other network. It is a system built using commodity hardware components, like any PC capable of running Linux, standard Ethernet adapters and switches. This makes the Beowulf cheap. But it is not a layman's job to implement a Beowulf cluster. Is implementation requires proper designing and configuration.

 The Khwarzimic Science Society has laid the stepping stone for Beowulf machines in Pakistan. Keeping abreast of the international wave towards high-performance computing using available resources is typically suited for a developing but progressive milieu like ours.

Introduction to Cluster Computing: What is a cluster; what alternatives does a cluster provide for high-performance computing and what are its contenders; the technologies, components and applications; supercomputer systems and issues; the IEEE Special Task Force on Cluster Computing; educational resources and Cluster Computing at the Khwarzimic Science Society. (Rajkumar Buyya, Mark Maker, Dan Hyde, edited by Bilal Muddassir).

Here is a list of books for reference on cluster computing.

 Special Thanks:

• Dr. Rajkumar Buyya, IEEE Computer Society Special Task Force on Cluster Computing, Monash International University, Australia.
• Homoeopathic Stores and Hospital, Lahore.

Mail the whole Cluster team / Mail the Web team