Multicast Optical Switched Network Core Node Structure

The core node is the most important device of optical switching networks. Its main function is to analyze,process and copy the optical signal which comes from the edge node and then transfer it to desire destination. therefore, the low time-delay,low power-loss,simple topology and nonblocking ability is essential to the core node. Recently, for the exponential growth of users and traffic of networks and lack of all-optical equipments, the key node will face lots of challenges, however, the core node is becoming more and more attractive because of its powerful processing ability and many amazing merits. In recent years, technology of the core node has made a mighty advance, and with progress of the technology of optical switching, the core node will get more and more accomplishments.Firstly, some classic architectures such as SPN,SPC,SPL are analyzed, and then, a space unit named MWC-MSD which has multicast capability is proposed. Three architectures of core node are presented, and the comparison of hardware complexity of them is given.Secondly, the three architectures are widesense-nonblocking which proved by mathematic means. then, the architecture is modified according to the principle of 3-stage clos networks, in the modified architecture, the middle switching stage is the AWGR. Then, the conditions when the modified architecture is strictly noblocking,rearrangeably nonblocking and widesense nonblocking is respectively given. 1,MMA is strictly nonblocking when m≥f+w-1;2,MMA is rearrangeable nonblocking when m≥max{f²,w}; 3,MMA is wide-sense nonblocking when m>2(f - 1)lgf / lglgf + (w - 1) (lgf)^1/2,f≥w m>(f - 1) [ lgf ] + 2(w -1),f < w。Thirdly, because of the high cost and technologic difficulties, using the LWC to substitute for the FWCs in MWC, some examples are given, and then the general conclusions is get, and it indicates that the measure is correct.Finally, the schedule algorithms Algorithms₁ and Algorithms₂ are presented, and the simulation graph is given with different traffic intensityρw and multicast fan-outθ. a algorithms which based on priority and its simulation graph are given. With the simulation, the Algorithms₂ performs well than Algorithms₁.The summary gives disadvantages of the scheme and the future work we should do.