Research Of Clustering And Routing Algorithm Of Hierarchical Optical Network

ASON is new generation optical transmission network which can intelligently and automatically fulfill the switching and connecting function of optical network. Routing is the key technology of ASON, and is important in dynamic route selection stage in ASON connection establishment. There are three routing technologies in ASON, and hierarchical routing is one of them.This work mainly focuses on hierarchical routing problem in ASON. The purpose of our research is to reveal the factors which influence the performance of hierarchical routing. And according to that we can design well performed hierarchical routing algorithm and clustering algorithm.We found that hierarchical clustering, static routing table and dynamic route selection imposed on blocking performance of optical hierarchical network. In particular, there are two kinds of influences, which are that lower domains shield route information from higher domains, and that inter-domain link has heavy load. We call them lower domain's shielding and heavy load of inter-domain link.Based on the routing and wavelength assignment technology of traditional flat network, we did research on fixed routing and fixed alternative routing, and found that these mature technologies could improve the routing performance of optical hierarchical network. Further more, we developed a novel routing algorithm, which run fixed alternative routing on inter-domain routing, and run fixed routing on inner-domain routing. So this routing algorithm could reduce the cost of installing fixed alternative routing on entire network. Through math analyze and simulation, we proved that this algorithm could get good performance gain by little cost.Through research of hierarchical routing, we found that there was little performance gain by using current routing technology. By contrast, the hierarchical structure had more obvious influence on routing performance of optical hierarchical network. So we did forward research on the influence of hierarchical structure on hierarchical routing.We abstracted two properties (the number of clusters, the variance of cluster size distribution) from hierarchical structure, and through which we proposed a property a (the total kind of inter-cluster traffic/the total kind of inner-cluster traffic). We used a to build connection between the static property (hierarchical structure) and dynamic property (blocking performance), and figured out how to do clustering could make hierarchical structure having small blocking probability. Later, we proposed a framework that could asses and evaluate the performance of hierarchical structure.At last, we developed a simulation platform providing convenient user interface to input topology, to do clustering, to develop hierarchical routing algorithm, and to run simulation. The result proved our conclusions.