近年來隨著網路的迅速發展，點對點(Peer-to-Peer)的檔案傳輸方法相較於傳統的主從式架構來說具有快速、穩定、以及節省資源等各項優點，因而逐漸取代傳統檔案傳輸的模式。BitTorrent就是根據點對點檔案傳輸方式所衍生出的一個應用程式，而在廣大的BitTorrent使用戶中，必定隱含許多資訊可提供給使用者，包括使用者的使用習性、所喜愛的檔案，及其他各項特徵。 在本篇論文中，我們設計了一個針對BitTorrent使用者的資料分析系統。再搭配資料倉儲及資料探勘技術，將資料有系統的整合與管理，而由此整合後的資料，我們可以迅速並且得到更具有代表性意義的資訊及分析以提供給使用者參考利用。 Recently, BitTorrent has emerged as a very popular and scalable peer-to-peer file distribution mechanism. It has been successful at distributing large file quickly and efficiently. With traditional client-server architecture expensive large server farms or mirroring are used to satisfy large number of requests. On the other hand, BitTorrent users act simultaneously as clients and servers. With large number users of BitTorrent, there should be much information about the user data flowing through network. The rich information may imply the users’ habitual behavior, data access pattern, interested file, and so on. The information is useful for general users, network managers, and designer, etc. Hence, we propose a framework on BitTorrent which combined with data warehouse and data mining techniques to offer an efficient and systematic analysis for users.
