摘要
本文提出了一种共享波长转换器和光纤延迟线的全光分组交换结构 ,与传统的设计方法和实验报道相比 ,该结构提供了更灵活的冲突解决措施 .为了获得与在交换系统输入级分别为每个波长信道提供波长转换器相等价的性能 ,本文给出了交换结构最多所需的转换器数量 .仿真结果表明 ,本文结构是兼顾交换系统体积、成本和性能等三方面较理想的折衷方案 .研究还表明 ,光纤数与每纤的波长数量之积愈大 ,本文结构所节约的转换器和延迟线数量将愈加可观 .
A novel all-optical packet switching node architecture is proposed,which is equipped with a set of fiber delay lines (FDLs) and tunable optical wavelength converters (TOWCs) shared among the input lines.As far as contention resolution is concerned,the developed architecture is more flexible and cost effective than the traditional design and the methods introduced in the literatures.To obtain an equivalent packet loss probability to the method that provides TOWCs for each input wavelength channel,we give the maximum number of converters required by the system.Simulation results show that the developed architecture can guarantee a good tradeoff in terms of system size,cost and performance.We also get the conclusion that with the increase of the product of the number of input/output fibers and the number of wavelengths,the number of TOWCs and FDLs needed will,being different from the traditional design,gradually decrease.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2004年第5期713-717,共5页
Acta Electronica Sinica
基金
国家自然科学基金 (No .69990 540 )
国家 863计划资助 (No .863 31 7 2 0 0 1AA1 2 1 0 73)
关键词
全光分组交换
光缓存
波长转换器
共享
分组丢弃率
all-optical packet switching
optical buffering
wavelength converter
sharing
packet loss probability
