Energy efficiency in P2P systems and applications

Abstract

This chapter presents a general taxonomy to classify possible approaches to the energy problem in peer-to-peer (P2P) systems and applications. It provides a classification of energy-aware P2P approaches into three main classes, based on the techniques they use, namely, sleep/wakeup, hierarchical, and resource allocation. The chapter surveys the main solutions available in the literature, focusing on two relevant classes of P2P protocols, namely, file-sharing/distribution protocols (e.g., BitTorrent and Gnutella) and epidemic P2P protocols. It compares, in terms of energy efficiency, the two main approaches to file sharing, that is, client-server and P2P. The chapter also presents some techniques for optimizing the energy efficiency in P2P file sharing. BitTorrent implements an unstructured overlay network customized for file sharing. Designing energy-efficient epidemic (gossip-based) protocols and services has become significant because of their wide usage in large-scale distributed systems. There exist two main classes of epidemic algorithms, namely, flat and hierarchical.