空间信息网络业务建模

Spatial Information Network Traffic Modeling

网络业务建模可用于进行业务的预测,对于卫星这种带宽受限的通信系统尤为重要。针对空间信息网络业务的特点,介绍了一种分形对数正态噪声(FLN)与Poisson模型相叠加的模型。该模型既具有自相似性也有短相关性。其中分形对数正态噪声(FLN)是分形高斯噪声(FGN)的转变,其统计特性可以根据流量和数据源特性进行任意的调节,比较精确灵活。Poisson模型作为最经典的业务模型,适用于短相关性,易于实现。

Network traffic modeling may be used for traffic prediction and is particularly important for satel- lite communication system with limited bandwidth. In accordance with to the characteristics of spatial infor- mation network traffic, a model based on superposition of FLN （fractional log-normal noise） and Poisson process is presented. This model is of self-similarity and short-range dependence. FLN is transformed from FGN and its statistical characteristics may be easily according to flow and data source properties, thus it is precise and flexible. Poisson process, as the most classic traffic model, is suitable for short-range de- pendence and easy to implement.