Most P2P systems divide a file into many pieces and fetch different pieces from many peers simultaneously. If one of the last few pieces is requested from an extremely slow peer, the overall download time may become excessive. The end-game mode was proposed to solve this problem. This method requests the last piece from all the connected peers at the same time. Unfortunately, the duplicated requests generate redundant traffic. This paper proposes a requesting method that does not increase overall download time as well as avoiding the redundant traffic. We also propose a combination of our requesting method and the end-game mode. The proposed methods are compared to the existing method with the end-game mode in simulations that use BitTorrent as a typical instance of the parallel retrieving P2P system. The results confirm that our requesting method can match the download times of the end-game mode while suppressing the redundant traffic. Our method enhances network performance by absorbing the difference in peer performance and providing steady download times without wasting traffic resources. Moreover, it is also confirmed that our combination method can distribute a file more quickly than other methods.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Junichi FUNASAKA, Akihiko ISHIZU, Kenji ISHIDA, "A File Fetching Method to Avoid Performance Deterioration on BitTorrent-Like P2P Networks" in IEICE TRANSACTIONS on Communications,
vol. E92-B, no. 8, pp. 2591-2599, August 2009, doi: 10.1587/transcom.E92.B.2591.
Abstract: Most P2P systems divide a file into many pieces and fetch different pieces from many peers simultaneously. If one of the last few pieces is requested from an extremely slow peer, the overall download time may become excessive. The end-game mode was proposed to solve this problem. This method requests the last piece from all the connected peers at the same time. Unfortunately, the duplicated requests generate redundant traffic. This paper proposes a requesting method that does not increase overall download time as well as avoiding the redundant traffic. We also propose a combination of our requesting method and the end-game mode. The proposed methods are compared to the existing method with the end-game mode in simulations that use BitTorrent as a typical instance of the parallel retrieving P2P system. The results confirm that our requesting method can match the download times of the end-game mode while suppressing the redundant traffic. Our method enhances network performance by absorbing the difference in peer performance and providing steady download times without wasting traffic resources. Moreover, it is also confirmed that our combination method can distribute a file more quickly than other methods.
URL: https://global.ieice.org/en_transactions/communications/10.1587/transcom.E92.B.2591/_p
Copy
@ARTICLE{e92-b_8_2591,
author={Junichi FUNASAKA, Akihiko ISHIZU, Kenji ISHIDA, },
journal={IEICE TRANSACTIONS on Communications},
title={A File Fetching Method to Avoid Performance Deterioration on BitTorrent-Like P2P Networks},
year={2009},
volume={E92-B},
number={8},
pages={2591-2599},
abstract={Most P2P systems divide a file into many pieces and fetch different pieces from many peers simultaneously. If one of the last few pieces is requested from an extremely slow peer, the overall download time may become excessive. The end-game mode was proposed to solve this problem. This method requests the last piece from all the connected peers at the same time. Unfortunately, the duplicated requests generate redundant traffic. This paper proposes a requesting method that does not increase overall download time as well as avoiding the redundant traffic. We also propose a combination of our requesting method and the end-game mode. The proposed methods are compared to the existing method with the end-game mode in simulations that use BitTorrent as a typical instance of the parallel retrieving P2P system. The results confirm that our requesting method can match the download times of the end-game mode while suppressing the redundant traffic. Our method enhances network performance by absorbing the difference in peer performance and providing steady download times without wasting traffic resources. Moreover, it is also confirmed that our combination method can distribute a file more quickly than other methods.},
keywords={},
doi={10.1587/transcom.E92.B.2591},
ISSN={1745-1345},
month={August},}
Copy
TY - JOUR
TI - A File Fetching Method to Avoid Performance Deterioration on BitTorrent-Like P2P Networks
T2 - IEICE TRANSACTIONS on Communications
SP - 2591
EP - 2599
AU - Junichi FUNASAKA
AU - Akihiko ISHIZU
AU - Kenji ISHIDA
PY - 2009
DO - 10.1587/transcom.E92.B.2591
JO - IEICE TRANSACTIONS on Communications
SN - 1745-1345
VL - E92-B
IS - 8
JA - IEICE TRANSACTIONS on Communications
Y1 - August 2009
AB - Most P2P systems divide a file into many pieces and fetch different pieces from many peers simultaneously. If one of the last few pieces is requested from an extremely slow peer, the overall download time may become excessive. The end-game mode was proposed to solve this problem. This method requests the last piece from all the connected peers at the same time. Unfortunately, the duplicated requests generate redundant traffic. This paper proposes a requesting method that does not increase overall download time as well as avoiding the redundant traffic. We also propose a combination of our requesting method and the end-game mode. The proposed methods are compared to the existing method with the end-game mode in simulations that use BitTorrent as a typical instance of the parallel retrieving P2P system. The results confirm that our requesting method can match the download times of the end-game mode while suppressing the redundant traffic. Our method enhances network performance by absorbing the difference in peer performance and providing steady download times without wasting traffic resources. Moreover, it is also confirmed that our combination method can distribute a file more quickly than other methods.
ER -