Structured peer-to-peer systems are designed for a highly scalable, self organizing, and efficient lookup for data. The key space of the so-called distributed hash tables (DHTs) is partitioned and each partition with its keys and values is assigned to a node in the DHT. For data retrieval however, the very nature of hash tables allows only exact pattern matches. We propose Cerco, a simple solution for the problem of range queries by employing a hierarchically structured P2P approach based on the principles of distributed hash tables. We show that a dynamic hierarchy of DHTs with on-demand classification of items can positively influence the response time of queries while maintaining lookup correctness
Published in:
Local Computer Networks, Proceedings 2006 31st IEEE Conference on
Date of Conference: 14-16 Nov. 2006