Concepedia

Abstract

Peer-to-peer backup systems provide mechanisms that allow participants to store their data cooperatively. When the proportion of unstable peers grows, these systems need to increase redundancy to maintain data availability. This redundancy could be reduced by storing data to more stable peers. In consequence, network traffic would be reduced and the global network¿s storage capacity would be increased. In this work we present a novel distributed backup system that improves data availability and network fairness. Using only local information our system creates groups of peers with similar stabilities. These groups are used to choose partners and to exchange data symmetrically with them. With this technique we guarantee that all peers obtain a data availability proportional to their stability and that all peers receive the same backup capacity than the disk space that they share.

References

YearCitations

Page 1