Making exploitation of peer heterogeneity as a first class citizen for resource discovery in peer-to-peer networks

Hung-Chang Hsiao, Chung Ta King, Shih Yen Gao

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Resource discovery in structured and unstructured peer-to-peer (P2P) systems have been extensively studied. Surprisingly, most of the systems are not designed to take advantage of node heterogeneity. In this paper, we propose a novel overlay called RATTAN, which serves as an underlay for unstructured P2P networks. RATTAN exploits the heterogeneity of nodes by structuring capable nodes as the core network of an unstructured P2P system. With RATTAN as the underlay, the scope of resource discovery in an unstructured P2P system can be maximal using a minimal number of messages. We evaluated RATTAN in simulation. The results show that (1) RATTAN is robust by exploiting redundant overlay links, and (2) the maximum bandwidth for protocol processing in a single RATTAN overlay is around 1 Mbits/sec, where 80% of nodes merely take 66 Bits/sec. We believe that a desktop machine equipped with an 100 Mbits/sec network interface is capable of processing 1 Mbits/sec of protocol operations. Peers that connect to the overlay via slow access networks, e.g. modems with 56 Kbits/sec, can accommodate the 66 bits/sec of overhead.

Original languageEnglish
Pages (from-to)952-961
Number of pages10
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3207
Publication statusPublished - 2004

Fingerprint

Resource Discovery
Peer to peer networks
Peer-to-peer Networks
Overlay
Exploitation
P2P Systems
Network protocols
Modems
Vertex of a graph
Processing
Interfaces (computer)
Telecommunication links
Bandwidth
Peer-to-peer Systems
Peer-to-peer (P2P)
P2P Network
Class
Simulation

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

@article{41627ad316234842bf5ea4a3397a444a,
title = "Making exploitation of peer heterogeneity as a first class citizen for resource discovery in peer-to-peer networks",
abstract = "Resource discovery in structured and unstructured peer-to-peer (P2P) systems have been extensively studied. Surprisingly, most of the systems are not designed to take advantage of node heterogeneity. In this paper, we propose a novel overlay called RATTAN, which serves as an underlay for unstructured P2P networks. RATTAN exploits the heterogeneity of nodes by structuring capable nodes as the core network of an unstructured P2P system. With RATTAN as the underlay, the scope of resource discovery in an unstructured P2P system can be maximal using a minimal number of messages. We evaluated RATTAN in simulation. The results show that (1) RATTAN is robust by exploiting redundant overlay links, and (2) the maximum bandwidth for protocol processing in a single RATTAN overlay is around 1 Mbits/sec, where 80{\%} of nodes merely take 66 Bits/sec. We believe that a desktop machine equipped with an 100 Mbits/sec network interface is capable of processing 1 Mbits/sec of protocol operations. Peers that connect to the overlay via slow access networks, e.g. modems with 56 Kbits/sec, can accommodate the 66 bits/sec of overhead.",
author = "Hung-Chang Hsiao and King, {Chung Ta} and Gao, {Shih Yen}",
year = "2004",
language = "English",
volume = "3207",
pages = "952--961",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Making exploitation of peer heterogeneity as a first class citizen for resource discovery in peer-to-peer networks

AU - Hsiao, Hung-Chang

AU - King, Chung Ta

AU - Gao, Shih Yen

PY - 2004

Y1 - 2004

N2 - Resource discovery in structured and unstructured peer-to-peer (P2P) systems have been extensively studied. Surprisingly, most of the systems are not designed to take advantage of node heterogeneity. In this paper, we propose a novel overlay called RATTAN, which serves as an underlay for unstructured P2P networks. RATTAN exploits the heterogeneity of nodes by structuring capable nodes as the core network of an unstructured P2P system. With RATTAN as the underlay, the scope of resource discovery in an unstructured P2P system can be maximal using a minimal number of messages. We evaluated RATTAN in simulation. The results show that (1) RATTAN is robust by exploiting redundant overlay links, and (2) the maximum bandwidth for protocol processing in a single RATTAN overlay is around 1 Mbits/sec, where 80% of nodes merely take 66 Bits/sec. We believe that a desktop machine equipped with an 100 Mbits/sec network interface is capable of processing 1 Mbits/sec of protocol operations. Peers that connect to the overlay via slow access networks, e.g. modems with 56 Kbits/sec, can accommodate the 66 bits/sec of overhead.

AB - Resource discovery in structured and unstructured peer-to-peer (P2P) systems have been extensively studied. Surprisingly, most of the systems are not designed to take advantage of node heterogeneity. In this paper, we propose a novel overlay called RATTAN, which serves as an underlay for unstructured P2P networks. RATTAN exploits the heterogeneity of nodes by structuring capable nodes as the core network of an unstructured P2P system. With RATTAN as the underlay, the scope of resource discovery in an unstructured P2P system can be maximal using a minimal number of messages. We evaluated RATTAN in simulation. The results show that (1) RATTAN is robust by exploiting redundant overlay links, and (2) the maximum bandwidth for protocol processing in a single RATTAN overlay is around 1 Mbits/sec, where 80% of nodes merely take 66 Bits/sec. We believe that a desktop machine equipped with an 100 Mbits/sec network interface is capable of processing 1 Mbits/sec of protocol operations. Peers that connect to the overlay via slow access networks, e.g. modems with 56 Kbits/sec, can accommodate the 66 bits/sec of overhead.

UR - http://www.scopus.com/inward/record.url?scp=35048855611&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35048855611&partnerID=8YFLogxK

M3 - Article

VL - 3207

SP - 952

EP - 961

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -