點對點(P2P，Peer-to-Peer)的檔案分享方式愈來愈流行，如ezPeer、BitTorrent…等，尤其以BitTorrent (BT)方式分享者，更是以大檔案分享為主，再加上其分享的原理為不分區域、遠近，大家隨機連線，容易造成資料重複地傳送，無可避免的，定會造成頻寬的嚴重浪費。特別是在區域網路(LAN)中，對外的頻寬常常是不夠用的，若再加上BT的使用，更容易造成網路的擁塞，甚至網路因此癱瘓而無法使用。因此，本文主要在探討於有使用私有IP(private IP)與防火牆(Firewall)的區域網路中使用BT軟體時，如何節省不必要之對外頻寬的浪費，進而當多人使用BT下載同一檔案時，能比BT原本的方法更快完成分享的動作！
因為分享者及追縱伺服器(Tracker)之埠號 (port number)並不固定，所以修改torrent的metainfo檔(.torrent)或利用IP名單做導向的方法是必要的，利用網路位址轉換(NAT)的功能，可將內部的使用者導向經過修改後且置於區域網路內的區域網路追蹤伺服器(LAN-Tracker)，由內部的區域網路追蹤伺服器給予下載同一torrent的下載者名單，這樣可將連線的對象限制為區域網路中的成員，進而節省對外的頻寬，並利用區域網路內高速的互傳速度，縮短分享的時間。而原始的檔案由本文中所設計的D-node這台快取機器代表區域網路中所有成員對外下載，因此同一份資料由區域網路外只需傳送一份至區域網路內。整體而言，不但大大地節省了不論是上傳或下載的頻寬，更加速了區域網路內分享同一檔案的速度！

In the file sharing systems, P2P(Peer-to-Peer) is more and more popular, like ezPeer and BitTorrent(BT). Especially the BitTorrent, it will make the network congested and even paralyzed because of it mainly shares large files and it’s sharing techniques. Especially between LAN(Local Area Networks) and the ISP network, the bandwidth is limited. If some users use BT program to download files inside the LAN that protected by private IP and firewall, it will cause the LAN in congestion. In this paper, we are going to try to save the bandwidth when using BT in a LAN and, more over, speed up the average download rate compared with the original download method.
Because the port numbers of the trackers and peers are not fixed, we have to use the blocking list in NAT or try to modify the .torrent file. This will make the peers inside the LAN connect to the LAN-tracker, instead of the original tracker outside the LAN. Making use of the LAN-tracker, we can limit the peers inside the LAN share files with each other without connecting to the peers outside the LAN. Only designated Peer designed in this thesis can connect to the peers outside the LAN, so it will prevent transferring the same data more than once. Finally, we can save the bandwidth and even speed up the average download time.

[1] Chonggang Wang, Bo Li, Peer-to-Peer Overlay Networks: A Survey, April 20, 2003
[2] ezPeer, http://www.ezpeer.com/
[3] Bram Cohen, Incentives Build Robustness in BitTorrent
[4] Indiana University Abilene Network, http://netflow.internet2.edu/weekly/
[5] BitComet, http://www.bitcomet.com/
[6] Bram Cohen, http://bitconjurer.org/
[7] Cache Logic™, http://www.cachelogic.com/
[8] Greek School Network, http://www.sch.gr/en/
[9] 陳宜凡, 應用於BT同儕通訊的通透閘道快取, July, 2005
[10] Asfandyar Qureshi, Exploring Proximity Based Peer Selection in a BitTorrent-like Protocol, May 7, 2004
[11]Autonomous System, http://en.wikipedia.org/wiki/Autonomous_system_(Internet) & RFC1930
[12] Sitaram Iyer Antony Rowstron Peter Druschel, Squirrel：A decentralized peer-to-peer web cache, the 21th ACM Symposium on Principles of Distributed Computing , July 2002
[13] The popularity of Gnutella queries and its implications on scalability, http://www2.cs.cmu.edu/~kunwadee/research/p2p/gnutella
[14] S. Patro and Y. C. Hu, Transparent Query Caching in Peer-to-Peer Overlay Networks, in Proceedings of the 17th International Parallel and Distributed Processing Symposium (IPDPS), 2003
[15] Chen Wang, Li Xiao, Yunhao Liu, and Pei Zheng, Distributed Caching and Adaptive Search in Multilayer P2P Networks, in Proceedings of the 24th International Conference on Distributed Computing Systems (ICDCS’04), 2004
[16] Adam Wierzbicki, Nathaniel Leibowitz, Matei Ripeanu, Rafał Wozniak, Cache Replacement Policies Revisited: The Case of P2P Traffic
[17]PeerAppliance™, White Paper:Effectively Managing Network Traffic through P2P Caching
[18]PeerCache™, http://www.joltid.com/index.php/peercache/
[19]Cachepliance, http://www.cachelogic.com/products/cachepliance.php
[20] bencode，http://blog.csdn.net/liquid_zigong/archive/2005/07/08/417257.aspx/ & http://en.wikipedia.org/wiki/Bencode
[21]PEX, http://en.wikipedia.org/wiki/Peer_exchange
[22]Squid, http://www.squid-cache.org/
[23]LRU(Least Recently Used), http://140.116.39.71/course/os/ppt/chap92.htm
[24]ipoque, http://www.ipoque.com/en/perf_detail.html#m1