摘要
为提高端到端可用带宽的预测精度,提出了一种基于自治域(AS)拓扑的可用带宽预测算法T-aware.该算法结合网络的AS拓扑信息,为符合超度量空间的节点和不符合超度量空间的节点分别选择测量点,对2类节点分别进行预测.在HP scalable sensing service数据集上的实验结果表明,T-aware的预测精度与PathGuru算法相比可提高10%以上.
To predict end-to-end available bandwidth accurately, a autonomous system (AS) topology-based available bandwidth prediction algorithm T-aware is proposed. T-aware employs the AS topology information to choose landmarks for ultrametric-constraint nodes and un-ultrametric-constraint nodes respectively, and then predicts the end-to-end available bandwidth accordingly. Experiments with HP scalable sensing service dataset show that the T-aware can predict end-to-end available bandwidth over ten percent precisely than the PathGuru algorithm.
出处
《北京邮电大学学报》
EI
CAS
CSCD
北大核心
2011年第3期66-70,共5页
Journal of Beijing University of Posts and Telecommunications
基金
国家自然科学基金项目(90304016)
江苏省自然科学基金项目(BK2009058
2010133)
关键词
可用带宽
算法
超度量空间
自治域路径
available bandwidth
algorithm
ultrametric space
autonomous system path