A traffic adaptive reconfiguration mechanism in multiwavelenth optical networks

Abstract

Multiwavelength optical network is promising to support the increasing bandwidth requirement of the Internet. $\emph{Reconfigurability}$ is one major feature in optical networks. This feature allows us to adapt the network topology to changed network state by reconfiguring the virtual topology with specific $\emph{reconfiguration policy}$. The objective of the reconfiguration that is carried out in response to changing traffic patterns is to improve network performance in terms of maximum throughput. However it is not the best way for maintaining or improving the network performance to reconfigure the virtual topology whenever traffic pattern changes. Therefore a frequency of the reconfiguration is controlled by a proper reconfiguration policy in a way to reduce the disruption of the connections. Virtual topology design and reconfiguration problem can be solved in mathematical formulation given many constraints and changed traffic matrix. But this optimization problem is known as NP-complete problem. Linear formulation problem by L (Linear Programming) package is optimized and exactly correct but its computation time is not tractable and takes so much time to apply dynamic traffic pattern change in real networks. Heuristic approaches can make solve large problem instances of the virtual-topology design and reconfiguration problem. Therefore we propose a $\emph{multi-stage reconfiguration}$ mechanism which is multi-stage decision process for reconfiguration instant. This heuristic counts the continual approximation by using predicted traffic with reward-cost function as reconfiguration policy. Therefore it is possible to adapt the virtual topology to the changed traffic positively. And it is based on $\emph{enhanced node exchange}$ algorithm which is proposed in this thesis also. This heuristic algorithm finds the best traffic adaptive virtual topology in a way to minimize the average weighted packet hop count. It is more progressive heuristic than convent...