K.: Popularity-Biased Random Walks for Peer-to-Peer Search under the Square-Root Principle (2006) [4 citations — 0 self]
Abstract:
The square-root principle is known to achieve low search time for peer-to-peer search techniques that do not utilize query routing indices (e.g., query flooding or random walk searches). Under this principle, each object is probed with probability proportional to the square root of its query popularity. Existing search methods realize the square-root principle by using either object replication or topology reconstruction, which may not be desirable for those applications with large, dynamic datasets and limited network bandwidth. We propose a new approach that uses popularity-biased random walks to achieve the square-root principle. With the guidance of the Metropolis algorithm, each step of the random walks is determined based on the content popularities of current neighbors. Compared to previous methods, our approach has comparable search performance at no cost of data movement or topology changes. 1.
Citations
| 122 | Replication strategies in unstructured peer-to-peer networks – Cohen, Shenker - 2002 |
| 57 | An introduction to Natural Computation – Ballard - 1997 |
| 30 | Exposé de la théorie des chaînes simples constantes de Markov á un nombre fini d’états. Mathématique de l’Union Interbalkanique, 2:77–105, π t , π – Doeblin |
| 12 | Biased Random Walks – Azar, Broder, et al. - 1992 |
| 8 | Quickly Routing Searches Without Having to Move Content – Cooper - 2005 |
