Routing on a curve

Relentless progress in hardware technology and recent advances in sensor technology, and wireless networking have made it feasible to deploy large scale, dense ad-hoc networks. These networks together with sensor technology can be considered as the enablers of emerging models of computing such as embedded computing, ubiquitous computing, or pervasive computing. In this paper, we propose a new paradigm called trajectory based forwarding (or TBF), which is a generalization of source based routing and Cartesian routing. We argue that TBF is an ideal technique for routing in dense ad-hoc networks. Trajectories are a natural namespace for describing route paths when the topology of the network matches the topography of the physical surroundings in which it is deployed which by very definition is embedded computing.We show how simple trajectories can be used in implementing important networking protocols such as flooding, discovery, and network management. Trajectory routing is very effective in implementing many networking functions in a quick and approximate way, as it needs very few support services. We discuss several research challenges in the design of network protocols that use specific trajectories for forwarding packets.