| Overlay multicast in mobile ad-hoc environments are finding newer applications everyday. Although overlay multicast is not as efficient as IP-based multicast, they have the advantage of being easy to implement. In many applications, some participating nodes might be members of more than one overlay tree or may wish to build a temporary tree in order to perform certain important tasks. The priority of these trees can be defined by the importance of the service. For the success of such applications, it is necessary that nodes belonging to more than one tree are smart enough to ignore incoming messages from members in low priority trees while they are listening to member from a higher priority tree. In this thesis, we present a Location-based Prioritized Overlay Multicast (L-POM) that builds priority trees with certain nodes carrying important tasks in overlay networks, and rearranges low priority trees using node location information whenever some nodes temporarily move to a high priority network. To obtain accurate location information, we design and implement an indoor location sensing system, LANDMARC. LANDMARC can improve the overall accuracy of locating objects by utilizing the concept of reference tags. Our experimental analysis demonstrates that active RFID is a viable and cost-effective candidate for indoor location sensing, and can be an effective support to L-POM. We further study the feasibility of L-POM by identifying a suitable unicast (ad-hoc) routing protocol, exploring to use location information to build more efficient trees, and studying the impact of density of wireless nodes, packet size, and fragmentation on performance. |
