Abstract
Peer-to-peer storage networks aim at aggregating the storage in today's resource-abundant computers to form a large, shared storage space. With the extreme heterogeneity in machines, networks and administrative organizations over which such networks are laid, peer-to-peer storage networks must be aware of the capabilities of the constituent components to leverage their resources, performance and reliability. This paper reports our design of such a peer-to-peer storage network called Tornado. Tornado is built on top of three concepts. The first is the virtual home concept, which designates each data item in Tornado to a virtual home instead of directly to a peer node. The second concept is node autonomy, which allows each peer node to determine the number of virtual homes it intends to host according to its computational capability. The third concept is quality classification, which classifies the peer nodes into ""good"" and ""bad"" according to their static and dynamic capabilities. Only ""good"" peers can host virtual homes, whereby leveraging the reliability and capabilities of these peers. When a good peer becomes overloaded, Tornado seeks another good but inactive peer to relieve the load. © 2003 Elsevier Inc. All rights reserved.