Date of Award

Spring 5-1999

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Science

Committee Director

Mohammad Zubair

Committee Director

Piyush Mehrotra

Committee Member

Kurt Maly

Committee Member

David Keyes

Call Number for Print

Special Collections LD4331.C65 L58

Abstract

Advances in computing and networking technologies are making large scale distributed heterogeneous computing a reality. Multi-Disciplinary Optimization (MDO) is a class of applications that is being addressed under this paradigm. It consists of multiple heterogeneous modules interacting with each other to solve an overall design problem. An efficient implementation of such an application requires scheduling heterogeneous modules (with different computing and disk 1/0 requirements) on a heterogeneous set of resources (with different CPU, memory, disk IO specifications).

Given a set of tasks and a set of resources, an optimal schedule of the tasks on the resources is very hard to compute (NP-Complete). In this study, we focus on scheduling of coarse grained tasks on heterogeneous resources. We propose two algorithms based on the classic static Critical Path Method (CPM). CPM has been suggested for homogeneous environments. We adapt this method for a heterogeneous environment. One of the proposed algorithms, named EA-CPM, maps a highest priority ready task to a processor that ensures its earliest assignment. The other algorithm, called EF-CPM, assigns a highest priority ready task to a processor that yields the earliest finishing time. We describe a series of qualitative, systematic numerical studies for evaluating algorithms. Overall, the performance of both algorithms is very promising. In particular, EF-CPM is close to the locally optimal solution obtained by the branch exhaustive search method.

Rights

In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).

DOI

10.25777/4et7-e348

Download

Share

COinS