P2P 檔案污染防治機制之模擬與分析

Abstract

為了遏止使用者於P2P檔案分享系統中擅自分享未經授權的電子檔案，唱片和電影等著作權公司放置了大量蓄意損毀的檔案於目前熱門的檔案分享系統，用以混淆使用者視聽，使之難以下載到正確的檔案。然而此一做法也漸漸被許多惡意的攻擊者用以降低檔案分享系統的效能。這種毀損檔案與正確檔案共存於系統中的現象被稱之為檔案汙染攻擊。檔案汙染攻擊不但降低了系統的效能，然而使用者下載到毀損檔案而重覆下載的行為也增加了許多不必要的網路傳輸流量。在本篇論文中，我們評估兩種具有代表性的檔案汙染防治機制抵抗目前盛行的兩種檔案汙染攻擊的程度。同時我們也探討使用者的行為對於檔案污染防治機制所帶來的影響。本篇的最後，我們提出了一個綜合兩種防治機制優點的方法，藉以創造一個更為有效率抵禦檔案汙染攻擊的P2P檔案分享系統。

In order to penalize unauthorized sharing of copyright-protected material in the Peer-to-Peer (P2P) systems, the music and film industries inject corrupted copies of the hit music and video clips to confuse the users and make it difficult for them to download a working copy of the desired file. Such practice is rather common and can be used by malicious users to compromise the data integrity of the P2P systems. This coexistence of corrupted and correct copies of files is referred to as file pollution in P2P systems. It does not only decrease the system performance, but also generates a large amount of traffic due to repeated downloads of the corrupted files. In this work, we evaluate the efficiency of two types of anti-pollution mechanisms on two popular pollution attacks. We also investigate how user slackness and awareness to the files influence the efficacy of the anti-pollution mechanisms. Finally, we propose a hybrid mechanism for more robust and efficient P2P system pollution protection.