臺灣博碩士論文加值系統

摘要

在眾多的peer-to-peer檔案分享系統之中，Bit Torrent以其檔案散佈的快速與頻寬資源高利用度等特色而獲得眾多使用者的愛好。但在使用Bit Torrent從事檔案下載的任務時，難免時常發生因為錯過了檔案分享的熱門時段或是缺乏擁有完整檔案的分享者，而造成的許多的peers皆因缺少檔案的部份而無法順利下載完成的情況；另一方面，檔案的下載者在透過tracker的仲介後，可能直接與檔案的發佈者進行溝通與檔案分享的動作。對於檔案的發佈者而言，其位址可能被有心的peer以觀察統計的方式查到，缺少了隱密性。
本論文主要在探討如何解決解決上述的二種Bit Torrent檔案分享系統使用上可能遭遇到的問題。在論文中，我們以Bit Torrent系統為基礎，實際設計和實作名為『主動式檔案擷取機制』與『匿名發佈機制』二種機制，分別對上述二種情況做判斷與處置。在雛形系統的測試裡，主動式檔案擷取機制能有效地減少因上述情況而發生的檔案無法順利下載之可能性，匿名發佈機制也能夠提供檔案發佈者一個有較佳隱匿性的選擇。除此之外，在實作機制中，我們也同時考慮了相容性的問題，使其能與原Bit Torrent系統一起進行檔案分享的動作。

Abstract

Because of faster file sharing speed and effective resource utilization, the Bit Torrent file distribution system becomes one of popular P2P file sharing systems. According to Bit Torrent protocol, peers can download the interesting file successfully if there exists one peer (or a set of peers) sharing the file at the moment. Consequently, peers might meet the situation such that files can not be downloaded completed due to lack of seeder (the peer own the complete copy of the file). Peers might miss the hot period of sharing the interesting file so that there is no seeder to share the file. Meanwhile, when peers download files, they communicate or exchange segments of the same interesting file with others, probably including the source provider. As a result, the source provider might leak its identity (the IP address) to those peers directly connecting with it.
In this thesis, we discuss how to solve the two problems above when using the Bit Torrent protocol. We design and implement two mechanisms: “Active File Access” and “Anonymous Publishing”, to handle the two problems respectively. The “Active File Access” mechanism can effectively reduce the failure probability of file download due to lack of seeders. The “Anonymous Publishing” mechanism enhances anonymity of file sharing for source providers in the Bit Torrent file distribution system. We have implemented a prototype system with Python, based on Bram Cohen’s source code. We show, from our prototype system, that the proposed mechanisms work well and are totally comparable with the Bit Torrent protocol.

目錄

第一章 序論........................................................1
1.1 前言........................................................1
1.2 研究動機....................................................3
1.3 論文架構....................................................3
第二章 P2P檔案分享系統相關研究....................................4
2.1 eMule的起源................................................4
2.1.1 eMule網路的一些基本概念..............................4
2.1.2 eMule網路的簡介......................................5
2.1.2.1 以eD2K為基礎的傳統伺服器.......................6
2.1.2.2 不需伺服器的Kademila網路.......................6
2.2 Bit Torrent的簡介...........................................7
2.2.1 Bit Torrent Protocol的介紹............................8
2.2.1.1 檔案分享的前置動作..............................8
2.2.1.2 tracker處理downloader的要求與回應.............10
2.2.1.3 Bit Torrent的 Peer Protocol....................13
2.3 討論.......................................................19
第三章 主動檔案擷取機制...........................................21
3.1 動機與待解決之問題.........................................21
3.2 系統機制設計...............................................23
3.2.1 tracker端............................................23
3.2.1.1 如何判斷downloader為seeder或leecher..........25
3.2.1.2 如何判斷檔案目前是否是處於可能無法順利完全下載.26
3.2.1.3 主動式詢問.....................................28
3.2.2 downloader端.........................................30
3.3 系統實作...................................................32
3.3.1 tracker端............................................32
3.3.2 downloader端.........................................34
3.3.3 與原Bit Torrent協定的相容性..........................37
第四章 匿名發佈機制...............................................39
4.1 動機與待解決之問題.........................................39
4.2 系統機制設計...............................................42
4.2.1 原始發佈者端.........................................42
4.2.2 代理發佈端...........................................43
4.3 系統實作...................................................44
4.3.1 原始發佈者端.........................................45
4.3.2 代理發佈端...........................................46
4.3.3 與Bit Torrent協定的共同運作.........................50
第五章 系統雛形與測試.............................................51
5.1 主動式檔案擷取機制雛形操作與測試...........................51
5.1.1 測試狀況與目標.......................................51
5.1.2 系統架設.............................................52
5.1.2.1 主要軟體設定...................................53
5.1.2.2 Web Server設定.................................53
5.1.3 系統測試.............................................54
5.1.3.1 Step 1: 啟動tracker............................54
5.1.3.2 Step 2: Peer 1與Peer 2連線至tracker..........55
5.1.3.3 Step 3: Peer 3連線至tracker...................61
5.2 匿名發佈機制雛形操作與測試.................................65
5.2.1 主要軟體設定.........................................65
5.2.2 系統架設.............................................66
5.2.2.1 測試條件設置與目標.............................67
5.2.2.2 Web Server與tracker的啟動.....................67
5.2.3 系統測試.............................................67
5.2.3.1 Step 1: Publish Server的啟動...................67
5.2.3.2 Step 2: Source Peer 與Publish Server的連線.....68
5.2.3.3 Step 3: Peer連線下載檔案.......................70
第六章 結論.......................................................74
參考文獻...........................................................75

圖目錄

圖 1.1：Server /Client架構圖.........................................1
圖 1.2：Peer-to-Peer架構圖..........................................2
圖 2.1：torrent範例檔案的內容.......................................9
圖 2.2：downloader所送出的要求訊息封包內容.........................12
圖 2.3：tracker所回應給downloader的訊息封包內容....................12
圖 2.4：Peer Protocol的流程說明封包(1)..............................14
圖 2.5：Peer Protocol的流程說明封包(2)..............................14
圖 2.6：Peer Protocol的流程說明封包(3)..............................15
圖 2.7：Peer Protocol的流程說明封包(4)..............................15
圖 2.8：Peer Protocol的流程說明封包(5)..............................16
圖 2.9：Peer Protocol的流程說明封包(6)..............................17
圖 2.10：Peer Protocol的流程說明封包(7)............................17
圖 2.11：Peer Protocol的流程說明封包(8)............................18
圖 2.12：Peer Protocol的流程說明封包(9).............................19
圖 3.1：原始Bit Torrent所可能遭遇到情形...........................22
圖 3.2：一個無法順利完成檔案下載的範例.............................24
圖 3.3：tracker所接受到的downloader連線至tracker時所送出的要求....25
圖 3.4：tracker判斷與處理的流程....................................28
圖 3.5：原始未修改的tracker所回覆的訊息封包........................30
圖 3.6：加入了主動式詢問機制後，tracker所回覆的訊息封包.............30
圖 3.7：一個completed.log範例檔案內容..............................31
圖 3.8：downloader處理ask_infohash欄位的流程.......................32
圖 3.9：在加入『主動式檔案擷取機制』之後可獲得改善之情況..............36
圖 4.1：tracker所回覆給downloader的訊息............................39
圖 4.2：原Bit Torrent機制可能遇到的情況.............................41
圖 4.3：匿名機制中原始發佈端發佈檔案的流程..........................43
圖 4.4：匿名機制中代理發佈端的處理流程..............................44
圖 4.5：加入了匿名發佈機制後的檔案分享情形..........................50
圖 5.1：IIS Server服務開放給user下載的檔案.........................54
圖 5.2：Active File Access Tracker的啟動.............................54
圖 5.3：Peer 1執行downloader的指令.................................55
圖 5.4：tracker所顯示Peer 1的downloader之要求訊息..................55
圖 5.5：Peer 1之downloader一直等待可供連線的peers而無法順利下載..57
圖 5.6：Peer 2執行downloader的指令................................58
圖 5.7：tracker所顯示Peer 2的downloader之要求訊息................58
圖 5.8：Peer 2之downloader一直等待可供連線的peers而無法順利下載....60
圖 5.9：Peer 3執行downloader的指令.................................61
圖 5.10：tracker所顯示Peer 3的downloader所送出要求訊息.............61
圖 5.11：tracker的console log訊息顯示..............................62
圖 5.12：Peer 3開啟downloader將檔案傳輸給Peer 1與Peer 2.............63
圖 5.13：Peer 1與Peer 2已經下載完成.................................64
圖 5.14：Peer 1的downloader已經下載完成所要的檔案了................64
圖 5.15：Peer 2的downloader也已經下載完成所要的檔案................65
圖 5.16：代理發佈伺服器的啟動.......................................68
圖 5.17：原始發佈者開啟Uploader程式連線至代理伺服器................69
圖 5.18：代理發佈伺服器收到Source Peer的連線........................69
圖 5.19：代理發佈伺服器開啟downloader進行檔案分享發佈的動作........70
圖 5.20：tracker顯示代理發佈伺服器所開啟的downloader連線之要求.....70
圖 5.21：Peer A開啟downloader程式連線到Tracker.....................71
圖 5.22：tracker的log訊息..........................................71
圖 5.23：Tracker回應給Peer A之downloader的訊息.....................72
圖 5.24：Peer A的下載情形...........................................72
圖 5.25：Peer A回應給tracker下載完成的訊息.........................73

表目錄

表 2.1：non-keepalive message的值與代表意義.........................15
表 3.1：tracker處置情況判斷........................................28
表 5.1：主動式檔案擷取機制測試用檔案資料............................52
表 5.2：主動式檔案擷取機制測試環境..................................52
表 5.3：匿名發佈機制測試用檔案資料..................................66
表 5.4：匿名發佈機制測試環境........................................66

參考文獻

[1] B. Cohen, “Incentives Build Robustness in BitTorrent”, May 2003, http://bittorrent.com/bittorrentecon.pdf
[2] D. Qiu and R. Srikant, “Modeling and performance analysis of bittorrent-like peer-to-peer networks”, In Proceedings of ACM Sigcomm (Portland, OR, Aug 2004).
[3] C. H. Ding, S. Nutanong, and R. Buyya, “Peer-to-Peer Networks for Content Sharing”, Technical Report, GRIDS-TR-2003-7, Grid Computing and Distributed Systems Laboratory, University of Melbourne, Australia, December 2003.
[4] P. Maymounkov and D. Mazieres, “Kademila: A peer-to-peer information system based on the xor metric”, In Proceedings of IPTPS02, Cambridge, USA, March 2002.
[5] S. M. Lui, and S. H. Kwok, “Interoperability of Peer-to-Peer Fiel Sharing Protocols”, ACM SIGecom Exchanges, Pages 25-33, Vol. 3, Issue 3, 2002.
[6] J. E. Berkes, “Decentralized Peer-to-Peer Network Architecture: Gnutella and Freenet”, University of Manitoba, Winnipeg, Manitoba, Canada, April 2003.
[7] K. Tutschku, “A Measurement-Based Traffic Profile of the eDonkey Filesharing Service”, in Proceedings of Passive and Active Network Measurement, 5th International Workshop, PAM 2004, Antibes Juan-les-Pins, France Apiil 19-220, 2004. (also appear in LNCS, Vol.3015/2004.)
[8] http://www.edonkey.com/

[9] Marshall Brain, Howstuffworks “How File Sharing Works”, http://computer.howstuffworks.com/file-sharing1.htm
[10] “Peer-to-Peer (P2P) and How Kazaa Works”, http://www.kazaa.com/us/help/glossary/p2p.htm
[11] http://www.bittorrent.com/index.html
[12] http://www.emule-project.net/
[13] http://www.python.org/
[14] http://sourceforge.net/project/showfiles.php?group_id=33044&package_id=25125
[15] Alan Kin Wah Yim, “Decentralized Media Streaming Infrastructure (DeMSI): A Peer-to-Peer Content Delivery Network”, Thesis of Master, Grid Computing and Distributed Systems Laboratory, Deparment of Computer Science and Software Engineering, University of Melbourne, Australia, November 2004.