基于数据中心流量特征的端到端流量估计算法

Traffic Estimation for Data Center Network Based on Traffic Characteristics

数据中心是云计算等大型分布式计算服务的基础,有效地设计与管理数据中心需要遵循数据中心网络的端到端流量特征。然而直接地测量网络的端到端流量需要耗费巨大的软件成本和硬件成本,并且由于数据中心网络结构的特殊性,传统的计算机网络采用的流量估计方法也无法适用于现有的数据中心网络。为解决以上问题,首先依据数据中心的资源分配和链路利用率情况提取出网络的粗粒度流量特征,在此基础上提出一种基于重力模型和网络层析技术的数据中心端到端流量估计算法。与现有的流量推理算法Tomogravity和ELIA在NS3搭建的不同规模的数据中心网络中进行性能对比,实验结果表明,所提算法能有效地利用提取出的粗粒度流量特征,在保证计算效率的前提下将计算准确度大幅提升,可满足当前数据中心网络实时获取端到端流量数据的需求。

Data center network（DCN）is the infrastructure of cloud computing and other distributed computing services.Understanding the characteristics of end-to-end traffic flows in DCNs is essential to DCN designs and operations.However,it is extremely difficult to measure the traffic flows directly.Due to the distinct structure of DCNs,the traditional traffic estimation method can not be applied to DCNs yet.To address this problem,we first extracted the coarsegrained traffic characteristics based on the user resource allocation and link utilization.And then an efficient traffic estimation algorithm was proposed for DCNs based on the gravity traffic model and network tomography.We compared our new proposal with two classical traffic inference algorithms Tomogravity and ELIA on different scale of DCNs.The results show that new algorithm outperforms the other two algorithms in both speed and accuracy.With the new method,the network managers can obtain the end-to-end traffic on DCNs in real time.