数据分组网中自相似业务模型的研究进展

Developments of modeling self-similar traffic indata packet networks

本文介绍了数据分组网络中业务模型的研究进展。数据分组网络中,传统的泊松或马尔科夫模型在描述网络业务的精确性方面有很大的不足,近年来发展的自相似(单分形)业务模型效果较好。最近,研究人员在实测网络业务数据的基础上提出的多分形模型,不但能很好地模拟网络业务的长相关性,还能表现其在小的时间尺度下的特性。本文简单介绍了单分形业务模型,然后对多分形业务模型进行了重点的阐述,对业务模型的研究进展做出了分析。

This paper introduces the developments of the traffic modeling of data packet network, in which the traditional Poisson and Markov models are not precise enough in modeling the data traffic for they could not catch the long-dependence of data traffic. The self-similar or fractal traffic models developed in recent years behave well to this point but not very well in characteristic of small time scale. Researchers have presented multi-fractal models based on data measured in real networks to make the traffic models more precisely. This paper mainly focuses on multi-fractal traffic models, by which means introduces the developments of traffic models in data traffic networks.