Premium
Optimal layer division for low latency in DHT‐based hierarchical P2P network
Author(s) -
Li Jun,
Li Cuilian,
Fang Zhaoxi,
Wang Haoyun,
Wu Yaohui
Publication year - 2016
Publication title -
international journal of network management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.373
H-Index - 28
eISSN - 1099-1190
pISSN - 1055-7148
DOI - 10.1002/nem.1922
Subject(s) - computer science , latency (audio) , computer network , distributed computing , scalability , distributed hash table , network layer , queuing delay , layer (electronics) , peer to peer , queueing theory , telecommunications , chemistry , organic chemistry , database
Summary Hierarchical architecture can significantly reduce the workload overheads and facilitate effective search in peer‐to‐peer (P2P) network by dividing the peers into two layers: a supernode layer and an ordinary node layer. System scalability is therefore improved, which is critical to the deployment of P2P applications, particularly for the P2P‐based Voice over Internet Protocol systems. However, the lack of appropriate size ratio of the supernode layer to the ordinary node layer makes the system search performance far from being optimal. To address the problem, the goal of this paper is to find the optimal layer ratio, such that lookup latency is minimized, subject to the capacity constraint of supernodes. With this particular objective, this paper first formulates a search delay model for the distributed hash table (DHT)‐based layered P2P architecture, taking account of two main sources of latency: DHT routing and queuing delay at the supernode. Then the optimal ratio between supernodes and ordinary nodes is found with the analytical model. Simulation and numerical results show that the obtained solution can improve the performance of a two‐layer P2P system in terms of lookup latency. Copyright © 2016 John Wiley & Sons, Ltd.