实时连续多媒体任务模型及调度算法  被引量:1

Real-time continuous multimedia task model and scheduling algorithm

在线阅读下载全文

作  者:阮俊波[1] 李红兵[1] 金惠华[1] 

机构地区:[1]北京航空航天大学计算机学院,北京100083

出  处:《北京航空航天大学学报》2005年第8期863-868,共6页Journal of Beijing University of Aeronautics and Astronautics

基  金:武器装备预研基金资助项目 (5 14 0 40 2 0 40 3HK0 14 9)

摘  要:传统周期任务模型过于简单,不能适应实时多媒体流任务的多种形式,以用户为中心的连续媒体的调度算法由于未能从全局上考虑任务的服务质量QoS(QualityofService)需求,因而调度效率较低.提出一种改进的以多媒体对象为中心周期任务模型,该任务模型概括了目前多媒体任务的基本表现形式.基于该任务模型,提出了一种以连续多媒体流的整个质量表现过程为中心的任务表现路径模型TPP(TaskPresentationPath) .在全面考虑表现路径中媒体流对象的全局死线、局部死线和任务连续失效率的基础上,给出了连续多媒体流的比例资源分配调度算法PST PP(ProportionalSharebasedonTaskPresentationPath) .实验结果表明了所提出的周期任务模型的合理性,调度算法能提供比传统分时策略以及EDF算法更好的QoS保证.The oversimplification of traditional periodic task model gives rise to poor flexibility, which can not adapt to multiple forms of real-time media stream applications. User-centered scheduling algorithm for continuous media brings in poor scheduling efficiency, for not considering QoS (quality of service)requirements completely. An improved periodic task model centered on multimedia objects was presented, and the model ould generalize the basic task representative forms of current multimedia applications on the whole. A model called TPP (task presentation path) for continuous media streams was proposed with focus on the whole qualities of presentation for continuous media stream. Considering the overall dead-line, partial periodic dead-line and sequential failure of multimedia stream object in the presentation path, a proportional share CPU scheduling called PSTPP(proportional share based on task presentation path) algorithm was brought forward for continuous multimedia streams. The experimental results show the rationality of models, and the scheduling algorithm can provide better QoS guarantees than the traditional timesharing and EDF strategies.

关 键 词:实时 错误 服务质量 连续媒体 任务调度 

分 类 号:TP37[自动化与计算机技术—计算机系统结构] TP316[自动化与计算机技术—计算机科学与技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象