PHYSICAL TOPOLOGY DESIGN AND ROUTING ALGORITHMS FOR DEGREE-CONSTRAINED FSO MESH NETWORKS

Abstract

Free-space optical (FSO) mesh networks are emerging as broadband communication networks because of their high bandwidth (up to Gbps), low cost, and easy installation. However, there are two existing problems in the deployment of FSO networks: the physical topology design problem, and the routing problem. This dissertation presents an algorithm for the physical topology design of FSO mesh networks in order to enhance network reliability under defined degree constraint of each FSO node. The methodology presented enlarges the minimum angle between adjacent links at each node. Simulation results show that, compared to other methods, the proposed algorithm not only provides higher connectivity and lower delay for FSO networks, but also makes the FSO networks so constructed more tolerant in a dynamically changing environment. Further, this algorithm is enhanced to include the 3-dimensional (3-D) space, where the heights of the FSO nodes are not identical. This enhancement will apply to FSO nodes in difficult terrains where it is not feasible or desirable to have the FSO transceivers on a plane.

This dissertation also addresses the routing problem in degree-constrained free-space optical (FSO) mesh networks. To solve the routing problem, four different routing algorithms are proposed. Their performances are evaluated through extensive simulations for a number of FSO mesh networks with different topologies and nodal degrees. The performance parameter against which these algorithms are evaluated is the mean end-to-end delay. The proposed least cost path (LCP) routing algorithm, which is based on minimizing the end-to-end delay, is considered as the bench mark. The performance of each of the other three algorithms is evaluated against the bench mark. The proposed minimum hop count with load-balancing (MHLB) routing algorithm is based on the number of hops between the source and the destination node to route the traffic. Simulation shows that the MHLB routing algorithm performs best in most cases compared with the other two. It results in minimum average delay and least blocked traffic.