Power-Efficient Ad Hoc Mobile Networking

点击翻译按钮获取中文摘要

Consider the problem of establishing a temporary network in a region hit by some natural disaster. Such a network would, for example, allow medics in the field to retrieve patient history from hospital databases, allow insurance companies to file claims from the field, etc. Ideally, setting up such a network should involve no more than equipping medics with PDAs (Personal Digital Assistants) that have radio capability and providing one or more ``gateways'' that connect this ad hoc network with the outside world. A fundamental problem in setting up such a network is routing packets between nodes in the ad hoc network. Why is this a difficult problem? The primary reason is that as nodes move about the network topology changes and thus routes need to be constantly updated. Furthermore, natural obstacles may have the effect of partitioning the network for brief periods of time. Thus, packets enroute to a destination that no longer lies in a partition may be dropped (even though the partition is short-lived). Finally, an overreaching constraint on any solution is the problem of very limited battery life at each mobile. Thus, while a specific protocol may route packets optimally, it may cause an inordinate amount of power drain causing mobiles to run out of power quickly. In this research, we will develop routing protocols that are power-efficient (i.e., the battery life of mobiles is maximized) while still providing near-optimal routing functionality. In addition, these protocols will support Quality of Service (QoS) guarantees by performing demand-based route discovery. Other protocol support for QoS, such as loss profiles for dynamic adaptation to bandwidth changes, will also be incorporated. Finally, the routing protocols will implement cache-and-forward packet routing to deal with short-lived network partitions. Papers and other documents related to this research project may be found at http://www.cs.sc.edu/~singh/adhoc.html.