Semantic partitioning of peer-to-peer search space

  • Authors:
  • Habib Rostami;Jafar Habibi;Emad Livani

  • Affiliations:
  • Computer Engineering Department, Sharif University of Technology, Tehran, Iran and IPM School of Computer Science, Tehran, Iran;Computer Engineering Department, Sharif University of Technology, Tehran, Iran and IPM School of Computer Science, Tehran, Iran;Electrical and Computer Engineering Department, University of Calgary, Canada

  • Venue:
  • Computer Communications
  • Year:
  • 2009

Quantified Score

Hi-index 0.24

Visualization

Abstract

Peer-to-peer (P2P) networks are beginning to form the infrastructure of future applications. Computers are organized in P2P overlay networks to facilitate search queries with reasonable cost. So, scalability is a major aim in design of P2P networks. In this paper, to obtain a high factor of scalability, we partition network search space using a consistent static shared upper ontology. We name our approach semantic partition tree (SPT). All resources and queries are annotated using the upper ontology and queries are semantically routed in the overlay network. Also, each node indexes addresses of other nodes that possess contents expressible by the concept it maintains. So, our approach can be conceived as an ontology-based distributed hash table (DHT). Also, we introduce a lookup service for the network which is very scalable and independent of the network size and just depends on depth of the ontology tree. Further, we introduce a broadcast algorithm on the network. We present worst case analysis of both lookup and broadcast algorithms and measure their performance using simulation. The results show that our scheme is highly scalable and can be used in real P2P applications.