Power efficient survivable routing with p-cycles

Abstract

Power awareness in networking has been a vital area of research in wireless networks but, until recently, has been largely ignored in wired networks. In wireless applications, the amount of power utilized by transmission is of vital importance since it will limit factors such as battery life and transmission range. In wired networks, the power issues of wireless networks do not arise since the wired networks receive their power from the power grid. However, the problem of operational costs and the environmental impact of wired networks have become increasingly important issues in recent years. This thesis proposes a power efficient routing scheme to address the environmental and operational cost issues. The operational costs of a wired network can be reduced by reducing the amount of power the network utilizes. The proposed power efficient routing scheme utilizes a demand prediction algorithm to determine a set of expected future traffic. The set of expected traffic is then assigned paths in the network using an energy efficient routing algorithm. The paths that are assigned to the predicted traffic are used to assign paths to the real traffic as it enters the network. By continuously updating the set of expected traffic, and the paths that are assigned to the expected traffic, the energy efficient routing algorithm can maintain an energy efficient routing solution over time, and thus, power efficiency is achieved.