基于成批排队的自相似业务性能分析

Performance Analysis of Self-Similar Traffic Based on Bulk Queue

自相似业务建模及性能分析是ATM领域当前的研究热点。讨论了当前排队模型的特点与不足,从分组层次上提出了一种新的模型。利用成批排队理论对系统的性能指标进行了分析,对模拟流量的自相似性及排队性能进行了实验仿真,研究了确定模型参数的方法,证实了理论结果的正确性。实验结果表明,从分组层次上建立的排队模型在性能上优于ON-OFF源层次上的模型。

In ATM field, current hotspot is the modeling and performance analysis of self-similar traffic. The characteristics and deficiencies of current queuing models are discussed and a novel model is brought forward in packet level. Its performance indices are analyzed through theory of bulk queues. Then, the self-similarity and queuing performance of the simulation traffic are experimentally researched. It is explored on how to define the model parameters and the theoretical results are proven to be true. The experimental results suggest that the queuing models established in packet level have better performance than those in ON-OFF sources level.