P2P Simulator for Queries Routing using Data Mining

TL;DR

This paper presents a P2P query routing simulator that leverages data mining, specifically Decision Trees, to predict suitable Super-Peers for queries, reducing network flooding and improving scalability.

Contribution

It introduces a novel integration of Decision Trees into P2P architectures for efficient query routing based on learned community models.

Findings

Predictive model reduces query flooding in P2P networks.

Knowledge-based routing improves robustness and scalability.

Decision Tree approach outperforms baseline flooding methods.

Abstract

Data mining is used to extract hidden information from large databases. In Peer-to-Peer context, a challenging problem is how to find the appropriate Peer to deal with a given query without overly consuming bandwidth. Different methods proposed routing strategies of queries taking into account the P2P network at hand. An unstructured P2P system based on an organization of Peers around Super-Peers that are connected to Super-Super-Peer according to their semantic domains is considered. This paper integrates Decision Trees in P2P architectures for predicting Query-Suitable Super-Peers representing a community of Peers, where one among them is able to answer the given query. In fact, by analyzing the queries' log file, a predictive model that avoids flooding queries in the P2P networks constructed by predicting the appropriate Super-Peer, and hence the Peer to answer the query. The proposed architecture is based on a Decision Tree (Base-Knowledge - BK). The efficiency of these architectures is discussed considering architecture without knowledge (Baseline) using only the flooding queries method to answer queries. The advantage of this knowledge based model is the robustness in Queries routing mechanism and scalability in P2P Network.