Wireless Mesh Networks (WMNs) are designed to provide broadband wireless services; however, capacities of WMNs are fundamentally limited in bandwidth because of the scarcity of RF spectral resources. A solution is to design a WMN using multiple wireless technologies. Since the optical spectrum remains unlicensed and under-utilized, we propose to augment RF WMNs with free space optical (FSO) systems. FSO links do not interfere with RF links and have higher capacity than RF links. In this paper, we study the throughput improvement of FSO-augmented RF WMNs. We address two questions. First, given a fixed number of FSO links, where should we install them to achieve the maximum improvement of network throughput for given traffic demands? Second, how do we route and schedule the flows in the hybrid FSO/RF network to achieve this throughput? We jointly formulate the above problems as one mixed integer linear programming problem. The results show that the throughput of the network increases dramatically by properly allocating FSO links. We also provide a computationally efficient algorithm to evaluate upper and lower bounds on the throughput of this hybrid FSO/RF network. The network administrator can use these bounds to determine the tradeoff between the number of FSO links (which are expensive) and the throughput improvement.
I. INTRODUCTIONWireless mesh networks (WMNs) have emerged to meet the growing demand for broadband wireless services. However, the performance of WMNs is limited by the cluttered RF spectrum. Free space optical (FSO) technology, using the unlicensed optical spectrum, is an alternative to RF technology for broadband wireless networks. FSO links can provide much higher data rate than RF links and do not interfere with RF links in the network. Among various wireless mesh network architectures [1], we are interested in infrastructure WMNs, as illustrated in Figure 1. Such WMNs consist of mesh routers (MRs) and mesh clients (MCs). The MRs in this infrastructure WMN form a backbone for MCs and are generally stationary. Thus, FSO technology, difficult to use in mobile environments but well adapted to stationary systems, can be added to this network to improve the network performance. Two major questions addressed in this paper are: (1) given a fixed number of FSO links, what is the maximum throughput of the network achievable by adding those FSO links? (2) how do we achieve this throughput by properly allocating the FSO links and choosing routing and scheduling schemes?There have been several studies on hybrid FSO/RF networks using different models based on their particular applications.