A Performance-Effective Pre-Scheduling Strategy for Grid Systems

碩士 === 中華大學 === 資訊工程學系碩士班 === 94 === Efficient task scheduling is an important issue to system performance in computational grid. To investigate this problem, the master slave paradigm is a good vehicle for developing tasking technologies of centralized grid system. In this thesis, we extend and i...

Full description

Bibliographic Details
Main Authors: CHEN TAI-LUNG, 陳泰龍
Other Authors: Ching-Hsien Hsu
Format: Others
Language:en_US
Published: 2006
Online Access:http://ndltd.ncl.edu.tw/handle/61014120986098269153
id ndltd-TW-094CHPI0392035
record_format oai_dc
spelling ndltd-TW-094CHPI03920352016-06-01T04:14:43Z http://ndltd.ncl.edu.tw/handle/61014120986098269153 A Performance-Effective Pre-Scheduling Strategy for Grid Systems 應用於主從式異質網格環境下的預先工作排程技術 CHEN TAI-LUNG 陳泰龍 碩士 中華大學 資訊工程學系碩士班 94 Efficient task scheduling is an important issue to system performance in computational grid. To investigate this problem, the master slave paradigm is a good vehicle for developing tasking technologies of centralized grid system. In this thesis, we extend and improve Beaumont’s scheduling strategies for dispatching tasks to heterogeneous processors in master slave environment. The main extension of this study is the consideration of heterogeneous communication overheads in grid system. One significant improvement of our approach is that average turnaround time could be minimized. The other advantage of the proposed method is that system throughput can be increased by dispersing processor idle time. Our proposed model can be applied on heterogeneous cluster systems within a grid environment in which the communication costs are various from clusters. To evaluate performance of the proposed techniques, we have implemented the proposed algorithms along with Beaumont’s method. The experimental results show that our techniques outperform Beaumont’s algorithm in terms of lower average turnaround time, higher average throughput, less processor idle time and higher processors’ utilization. Ching-Hsien Hsu 許慶賢 2006 學位論文 ; thesis 38 en_US
collection NDLTD
language en_US
format Others
sources NDLTD
description 碩士 === 中華大學 === 資訊工程學系碩士班 === 94 === Efficient task scheduling is an important issue to system performance in computational grid. To investigate this problem, the master slave paradigm is a good vehicle for developing tasking technologies of centralized grid system. In this thesis, we extend and improve Beaumont’s scheduling strategies for dispatching tasks to heterogeneous processors in master slave environment. The main extension of this study is the consideration of heterogeneous communication overheads in grid system. One significant improvement of our approach is that average turnaround time could be minimized. The other advantage of the proposed method is that system throughput can be increased by dispersing processor idle time. Our proposed model can be applied on heterogeneous cluster systems within a grid environment in which the communication costs are various from clusters. To evaluate performance of the proposed techniques, we have implemented the proposed algorithms along with Beaumont’s method. The experimental results show that our techniques outperform Beaumont’s algorithm in terms of lower average turnaround time, higher average throughput, less processor idle time and higher processors’ utilization.
author2 Ching-Hsien Hsu
author_facet Ching-Hsien Hsu
CHEN TAI-LUNG
陳泰龍
author CHEN TAI-LUNG
陳泰龍
spellingShingle CHEN TAI-LUNG
陳泰龍
A Performance-Effective Pre-Scheduling Strategy for Grid Systems
author_sort CHEN TAI-LUNG
title A Performance-Effective Pre-Scheduling Strategy for Grid Systems
title_short A Performance-Effective Pre-Scheduling Strategy for Grid Systems
title_full A Performance-Effective Pre-Scheduling Strategy for Grid Systems
title_fullStr A Performance-Effective Pre-Scheduling Strategy for Grid Systems
title_full_unstemmed A Performance-Effective Pre-Scheduling Strategy for Grid Systems
title_sort performance-effective pre-scheduling strategy for grid systems
publishDate 2006
url http://ndltd.ncl.edu.tw/handle/61014120986098269153
work_keys_str_mv AT chentailung aperformanceeffectivepreschedulingstrategyforgridsystems
AT chéntàilóng aperformanceeffectivepreschedulingstrategyforgridsystems
AT chentailung yīngyòngyúzhǔcóngshìyìzhìwǎnggéhuánjìngxiàdeyùxiāngōngzuòpáichéngjìshù
AT chéntàilóng yīngyòngyúzhǔcóngshìyìzhìwǎnggéhuánjìngxiàdeyùxiāngōngzuòpáichéngjìshù
AT chentailung performanceeffectivepreschedulingstrategyforgridsystems
AT chéntàilóng performanceeffectivepreschedulingstrategyforgridsystems
_version_ 1718286991542452224