基于自相似业务流的AOS延时累积调度算法  

Scheduling algorithm of delay accumulated adaptive polling based on AOS self-similar traffic

作　　者：赵运弢[1,2] 冯永新[1] 刘恒驰 刘猛[1] 

机构地区：[1]沈阳理工大学信息科学与工程学院,辽宁沈阳110159 [2]南京理工大学自动化学院,江苏南京210094

出　　处：《系统工程与电子技术》2015年第2期417-422,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(60802031;61101116;61471247);新世纪优秀人才支持计划(NCET-11-1013);辽宁"百千万人才工程"培养项目;沈阳理工大学2014年度重点学科;重点实验室开放基金资助课题

摘　　要：针对自相似业务流量下的高突发性及重尾性所引起的空间数据系统调度性能下降问题,分析了高级在轨系统(advanced orbiting system,AOS)虚拟信道存取(virtual channel access,VCA)子层调度策略以及现有基于短相关模型调度算法的不足,引入Hurst参数、紧迫度、流量离差、成帧时间因子等权值参量,提出一种基于延时累积的自适应轮询调度(scheduling of delay accumulated adaptive polling,SDAAP)算法,通过自适应改变延时阀值因子实现多业务的差异化调度,从而优化AOS虚拟信道服务质量及调度性能。采用多信源重尾分布的ON/OFF流量分布模型进行仿真验证,实验结果表明,针对自相似业务流,SDAAP算法在溢出率、平均延迟等方面优于AOS固定阀值和等时调度算法。In order to solve performance degradation for advanced orbiting system （AOS） space data system scheduling caused by high burst and heavy tailed nature of self-similar traffic, the existing problems of the AOS virtual channel access （VCA） layer scheduling strategy and the short correlation model of the scheduling algo- rithm are analyzed. A novel scheduling algorithm based on AOS delay accumulated adaptive polling （SDAAP） is proposed. Based on Hurst parameters, urgency, flow rate deviation, and framing time factor, the novel schedu- ling algorithm adaptively change the delay threshold factor to realize multi-service by＇ different operation meth- ods. The SDAAP algorithm optimizes the AOS virtual channel service quality and scheduling performance. With heavy tailed distribution of the ON/OFF traffic model, the experimental results show that, for the AOS self-similar traffic, the SDAAP algorithm executes more well in terms of the overflow rate and average delay than the AOS fixed threshold and the equal time scheduling algorithm.

关 键 词：高级在轨系统 自相似 重尾性 虚拟信道 调度 

分 类 号：TN919[电子电信—通信与信息系统]