随着全光接入网的不断发展,接入速率不断提高,对于边缘节点发送来的高速数据业务也必须进行流量梳理。为了适应未来全光高速网络接入的需求,基于光纤延迟线(Fiber delay lines,FDLs)结构,构造了一种简单高效的全光包交换网络边缘汇聚节点。并针对光线延迟线结构,提出一种新的主动缓存队列管理模型来分析节点性能,得到了与以往研究不同的结果。同时在对节点性能分析的基础上,提出一种新的FDLs的排列方式,在缓存深度相同的情况下,大大提高了FDLs的效率,实现了对到来的具有自相似性的数据业务的合理调配。 With the continuous development of all-optical access network, the access rate continues to increase, and traffic must be sorted for high-speed data services sent by edge nodes. In order to meet the demand of all-optical high-speed network access in the future, a simple and efficient all-optical packet switched network edge aggregation node is constructed based on the structure of Fiber Delay lines (FDLs). Aiming at the structure of light delay line, a new active buffer queue management model is proposed to analyze node performance. Different results are obtained from the previous studies. At the same time, based on the analysis of the node performance, a new arrangement of FDLs is proposed. With the same cache depth, the efficiency of FDLs is greatly improved and a reasonable deployment of self-similar data services is achieved .