大容量相干无源光网络的缓存机制研究

Research on caching mechanism of large capacity coherent passive optical network

针对当前大容量相干无源光网络节点持续传输数据流,大幅度提升网络支出与终端能耗,导致信息交互率较低、平均传输时间较长的问题。提出了大容量相干无源光网络的终端缓存机制。分析大容量相干无源光网络节点信息传输流程,通过求得信息交互率,结合多副本传输模型,获取优先级排序,采用ACK数据包清除冗余信息,利用Spray and Wait算法,设置副本可访问数量阈值,通过节点传输带宽与缓存空间特点,获得代替信息的理想架构。实验结果表明,所提方法的信息交互率较高,始终保持在90%左右,且平均传输时间仅为5 s,使得网络整体性能大幅度提高。

Aiming at the current large-capacity coherent passive optical network nodes was continuously transmitting data streams,significantly increasing network expenditures and terminal energy consumption,leading to the problem of low information exchange rate and long average transmission time. A terminal caching mechanism for large-capacity coherent passive optical networks is proposed. Analysing the information transmission process of large-capacity coherent passive optical network nodes,we obtain the information exchange rate by combining multiple copy transmission models. We use ACK packets to obtain priority ranking as well as clear redundant information and use the Spray and Wait algorithm to set the copy to be accessible. The number threshold,through the characteristics of node transmission bandwidth and cache space,obtains an ideal framework for replacing information. Experimental results show that the information exchange rate of the proposed method is relatively high,which is always maintained at about 90%,and the average transmission time is only 5 s,which dramatically improves the overall performance of the network.