计算机科学
計算機科學
계산궤과학
COMPUTER SCIENCE
2009年
11期
143-147,185
,共6页
网格工作流%可靠性%仿真工具
網格工作流%可靠性%倣真工具
망격공작류%가고성%방진공구
Grid workflow%Dependability%Simulation
网格工作流作为综合利用网格资源求解问题的"编程"技术已得到广泛应用,其可靠性保障研究得到越来越多的关注.然而,面对网格环境固有的复杂性和不确定性,如何时可靠性保障方法有效、方便地进行评测,是一个有待深入探索的挑战性问题.以VINCA网格工作流为参考原型,对影响其可靠性的因素进行了系统的分析,抽象出了网格工作流系统组件模型和可靠性属性模型,并基于GridSim实现了一个可配置、易扩展的网格工作流可靠性仿真工具包--VINCASim.此工具可以通过配置方便地模拟节点失效、工作流引擎失效、网络连接失效、流程执行异常等出错情形,提供扩展接口支持以编程的方式引入各种可靠性保障方法,为评测不同方法的有效性提供可控的、可重复的实验平台.通过场景示例说明了该工具的扩展性和易用性.
網格工作流作為綜閤利用網格資源求解問題的"編程"技術已得到廣汎應用,其可靠性保障研究得到越來越多的關註.然而,麵對網格環境固有的複雜性和不確定性,如何時可靠性保障方法有效、方便地進行評測,是一箇有待深入探索的挑戰性問題.以VINCA網格工作流為參攷原型,對影響其可靠性的因素進行瞭繫統的分析,抽象齣瞭網格工作流繫統組件模型和可靠性屬性模型,併基于GridSim實現瞭一箇可配置、易擴展的網格工作流可靠性倣真工具包--VINCASim.此工具可以通過配置方便地模擬節點失效、工作流引擎失效、網絡連接失效、流程執行異常等齣錯情形,提供擴展接口支持以編程的方式引入各種可靠性保障方法,為評測不同方法的有效性提供可控的、可重複的實驗平檯.通過場景示例說明瞭該工具的擴展性和易用性.
망격공작류작위종합이용망격자원구해문제적"편정"기술이득도엄범응용,기가고성보장연구득도월래월다적관주.연이,면대망격배경고유적복잡성화불학정성,여하시가고성보장방법유효、방편지진행평측,시일개유대심입탐색적도전성문제.이VINCA망격공작류위삼고원형,대영향기가고성적인소진행료계통적분석,추상출료망격공작류계통조건모형화가고성속성모형,병기우GridSim실현료일개가배치、역확전적망격공작류가고성방진공구포--VINCASim.차공구가이통과배치방편지모의절점실효、공작류인경실효、망락련접실효、류정집행이상등출착정형,제공확전접구지지이편정적방식인입각충가고성보장방법,위평측불동방법적유효성제공가공적、가중복적실험평태.통과장경시례설명료해공구적확전성화역용성.
As a special kind of "programming" technology for constructing problem-solving applications on the basis of grid resources,grid workflow has been widely applied.Methodologies for ensuring ddependability of grid workflows have attracted attention.However, it remains a challenge how to evaluate the effectiveness of these methodologies due to the complexity and uncertainty of grid environments.Based on VINCA grid workflow, key factors that affect the depen dability were systematically analyzed, and a general component model and dependability attributes model for grid work-flow systems were established.A configurable and extensible simulation toolkit called VINCASim for evaluating the de-pendability of grid workflows was developed based on GridSirn.The toolkit can simulate various failures raised from grid nodes, workflow engines, network connection and workflow execution in a configurable manner, and supports incor-porating different dependability ensuring methods programmatically.Thus, a controllable and repeatable experiment platform was provided for evaluating different methods.The usability was demonstrated by a use case scenario.