摘要
根据视频数据流的长程相关性特点,提出一种以太无源光网络中面向实时可变比特率视频的动态带宽分配算法.该算法利用自适应线性单步预测器对未来到达的平均视频流量进行预测,实时预测结果用于对下一个发送时隙的计算,使得带宽分配算法能够充分满足实时视频流量需求.仿真实验表明,所提出的算法能够准确地对实时可变比特率视频流量进行预测,有效地降低排队时延,减小队列长度并提高带宽利用率.
Based on the correlation structure of real-time variable bit rate video traffic, this paper proposed a dynamic bandwidth allocation (DBA) algorithm with adaptive linear traffic prediction in Ethernet passive optical network (EPON). A one-step adaptive linear predictor is used to predict the coming video traffic from users, and then the predicted results are used to help DBA accurately distribute bandwidth to satisfy users' needs. The simulation results show that this algorithm can significantly improve the quality-of-service (QoS) performances in terms of low delay, small buffer and high utilization.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2007年第8期1383-1387,共5页
Journal of Shanghai Jiaotong University
关键词
以太无源光网络
可变比特率视频业务
自适应线性流量预测
动态带宽分配算法
ethernet passive optical network (EPON)
variable bit rate video (VBR)
adaptive linear prediction
dynamic bandwidth allocation (DBA)