A new taxonomy for parallel Java models applied on cluster computing

  • Authors:

  • Mohamed Galal;Hesham Eldeeb;Salwa Nassar

  • Affiliations:

  • National Center for Examination and Educational Evaluation, Egypt;Electronic Research Institute, Egypt;Electronic Research Institute, Egypt

  • Venue:
  • ISTASC'04 Proceedings of the 4th WSEAS International Conference on Systems Theory and Scientific Computation
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

The efficient Java parallel programming models that make it easier for parallel programmers to synchronize and exchange data between processors in network of PCs cluster with different levels of parallelism, are needed. A cluster of PCs provides a scalable parallel hardware platform for high performance computing. Moreover, Parallel and distributed computing on a cluster of PC's is being increasingly applied to a variety of large size computational problems. A huge computational infrastructure can be presented by building a Java programming with its advantages on cluster computing which is characterized by its scalability, flexibility, and improvement of cost and time. In this paper, we propose a new taxonomy that is considered useful for describing the different models of parallelism that can be used for writing a parallel Java programs. Firstly, we present the taxonomy models including usage, advantages, and disadvantages of each model. Secondly, we provide a set of criteria to be used as a reference for evaluation and comparison of each model. In addition, we describe the design and the implementation of three parallel models and extract their features to make it easy for the parallel programs developers to choose appropriate Java model for there application.