一种新的光突发交换网络业务流的小波模型

New model of OBS network traffic based on wavelet

最近大量研究发现实际业务流表现出分形特性,这对业务流的性能产生较大影响。针对光突发交换(OBS)网络边缘节点的汇聚业务流,首先在多重分形条件下利用小波变换提出新的模型。与普通分形模型不同之处,该模型不仅考虑小波变换中各尺度因子的优化,而且充分考虑了实际网络环境(如缓冲区、利用率)等因素。同时经过实验证明业务流性能并非仅由业务流的相关性决定,而且与网络环境有较大的关系,在一定程度上甚至超过了相关性的影响。最后通过对业务流进行的尺度刻画和性能评价,验证了模型的有效性。

Recently the studies have found that real network traffic has the fractal characteristic, and it has great impact on the performance of traffic. With the convergence traffic of 0BS network edge node, the author proposed a new model based on the wavelet transform under the muhi-fractal condition. Different from the common fractal model, the new model considered not only the optimal of the scale factor in wavelet transform but also real network environment （ such as buffer size, utilization and so on）. Experimental results prove that the performance of traffic does not just relate with correlation, but also with the network environment. Finally, scale depiction and performance evaluation of traffic verify the effectiveness of the model.