Energy conscious communication protocols

Wireless networks, in the form of cellular, ad hoc and sensor networks, have become an integral part of modern society. A common thread limiting the true potential of these networks is their dependence on limited energy supplies. In this dissertation, we have explored the use of energy conscious communication protocols to extend network and node lifetimes. We argue that an awareness of application performance requirements at various layers of the protocol stack can lead to efficient use of battery resources.; We begin by addressing the problem of energy efficient packet scheduling in a wireless environment. Our objective is to design a transmission schedule that maximizes battery lifetime subject to some delay constraints. To achieve this, we exploit two previously unconnected ideas: (i) Channel coding can be used to conserve energy by transmitting at reduced power levels over longer durations; (ii) Electro-chemical mechanisms in batteries allows them to recover energy during idle periods. We develop a constrained dynamic programming framework to merge these two ideas and show that energy aware scheduling strategies can result in significant energy savings.; Caching is a popular strategy in wireline networks to reduce delay latency. In wireless networks the advantages of caching are offset by increased energy expenditure. We consider a network where time critical information needs to be disseminated at regular intervals. By comparing energy and delay costs, we propose a near optimal caching strategy. We have devised a distributed asynchronous caching strategy for implementing our caching strategy.; We also explore the energy savings possible by switching off a wireless node when it has no traffic to serve. While sleeping, the node consumes very little energy. The node wakes up periodically to serve incoming traffic from a base station. By sleeping for longer periods, the node conserves energy. However, packets get queued up at the base station resulting in larger delays. This gives rise to a trade-off between energy consumption and delay. We show that knowledge of the statistical properties of incoming traffic can be used to devise energy efficient wakeup strategies.