【作者】 吕雁飞； 武珊珊； 谷峪； 许嘉； 于戈；

【Author】 Lü Yanfei~1,Wu Shanshan~2,Gu Yu~2,Xu Jia~1,and Yu Ge~2 1(College of Software,Northeastern University,Shenyang 110004) 2(College of Information Science and Engineering,Northeastern University,Shenyang 110004)

【机构】 东北大学软件学院； 东北大学信息科学与工程学院；

【摘要】 数据流应用往往要求数据流上的连续查询能够满足用户期望的查询服务质量,以保证查询结果的有效性.针对数据流上的统计性QoS保证问题进行了研究,给出了统计性QoS的定义,并提出了一种基于网络演算的统计性QoS建模方法,使用随机到达曲线和服务曲线刻画了统计性查询QoS.在此基础上,提出了能够保证统计性QoS的SQED调度算法和相应的可调度性验证方法并给出了证明.实验表明,提出的统计性QoS保证策略能够有效地为数据流上连续查询提供统计性的QoS保证.更多还原

【Abstract】 Stream-based applications usually require that continuous queries over data streams meet certain expected query QoS,in order to make the results useful.The issue of statistic QoS guaranteeing over data streams is discussed in this paper.The concept of statistic QoS is defined,and a modeling approach based on network calculus for statistic QoS is proposed,in which statistic QoS can be properly described with stochastic arrival curves and service curves.On this basis,a statistic QoS guaranteeing scheduling algorithm named SQED,together with the corresponding schedulability verification method,is proposed and proved. Finally,sophisticated experiments show that the proposed approach offers statistic QoS guarantee to continuous queries over data streams efficiently.更多还原