Creating additional internet gateways for wireless mesh networks and virtual cell implementation using dynamic multiple multicast trees

Wireless Mesh Networks (WMNs) have become an important field of research in providing wireless Internet access at a lower cost and gained a lot of enthusiasm over the last few years. Although WMN exhibits attractive features, its multi-hop nature of communication limits the capacity that can be supported for any application. This limitation in capacity is seen as a major obstacle to large scale deployment of WMNs. The number of hops from Mesh Routers (MRs) to an Internet Gateway (IGW) plays an important role in determining the performance of a WMN. A recent patent has introduced a mechanism of using an existing Femtocell (FC) as an additional potential IGW so that the performance of a WMN could be enhanced. Femtocell technology is an emerging scheme introduced to increase the coverage inside the buildings where cellular coverage is usually quite low. FC technology also tries to route some portion of the cellular traffic going towards macro stations through Digital Subscriber Links (DSL), thus increasing the capacity of the macro station. Open Access, Closed Access and hybrid Access are three major types of FCs available. Open Access and Hybrid Access FCs let anyone around their coverage area to connect to them. Such an integration of a WMN with FCs enables an increase in WMN's overall capacity. But, due to FCs' unpredictable operating times and uninformed disconnections, MRs require reliable and efficient schedules for switching between available FCs. This has not been considered in the patent. The switching can be done in a push-based preemptive or a pull-based non-preemptive manner. In this dissertation, we formulate both of these switching schemes for a WMN-FC integrated network as approximate statistical models based on a Markovian process. We also determine a switching schedule of FCs for each of MR based on the Reliable Uncapacitated Facility Location (RUFL) problem. We extend an existing RUFL problem to incorporate dynamic operating nature of multiple FCs which possesses dynamic available/unavailable patterns as additional potential IGWs. Extensive simulations are carried out to validate our proposed statistical models and establish the performance of these switching schemes. Femtocells belong to special class of wireless networks known as Small Scale Networks (SCNs) consisting of a number of low range base stations. These SCNs are becoming increasingly popular among service providers as well as end customers. One major obstacle for SCNs is the management of the mobile users, resulting in frequent handovers which drastically reduce the overall performance. In this dissertation, we also focus on providing a framework that reduces the number of handovers using Dynamic Virtual Cells by utilizing Multiple Multicasting Trees. Members of a virtual cell are defined by the multicast tree membership. Virtual cells can move with the mobile users and dynamically grow or shrink depending on the performance. Handovers occur only between different virtual cells belonging to different multicast trees, thus minimizing effective number of handovers. After summarizing our research results, we discuss many potential extensions of our work and briefly discuss their feasibility.