Self-organizing, self-healing wireless networks

Abstract

Self-organizing, self-healing networks, also called "ad hoc" networks, are perhaps the most exciting new trend in networking research and development. Every node in such a network has sufficient intelligence to continuously sense and discover other nearby nodes, dynamically determine the optimal path for forwarding data packets from itself hop by hop through the network to any other node in the network, and automatically heal any ruptures in the network fabric that are caused by ongoing movement of the nodes themselves, changes in RF propagation, destruction of nodes, etc. In essence; one need merely launch such radio nodes into some space and they will not only organize themselves into a network but also adapt continuously to changes in the network's connectivity. The field is currently evolving at "Internet speed" - the first large-scale systems are now being fielded into the USA military and yet many of the field's research fundamentals are at present only very poorly understood. This paper begins with an example that introduces some of the fundamental issues for ad hoc networks (hidden terminals, channel access, mobility, scalability, and power constraints). It then discusses the two dominant approaches in current ad hoc networking technologies, "proactive" and "on-demand" path discovery, and compares the strengths and weaknesses of each approach. The paper then presents two brief case studies with which the authors were intimately involved - of a 400-node, terrestrial, vehicle-based system and of a high-capacity voice and data network deployed among drone aircraft - and concludes with highlights of the current research topics in the field.