在同儕網路上以可用頻寬量測機制偵測自私使用者之研究

近年來，同儕網路的相關軟體在今日的 Internet 上越來越受歡迎。然而，同儕網路天生就有使用者自私而影響效能的問題。自私的使用者為了增進他們自身的利益，常常會策略性的欺騙與他們互動的使用者。這些自私的行為會降低系統的效能以及增加系統的不穩定。因此，我們想研究使用者惡意降低流速的問題。我們提出一個可以在同儕網路可以降低使用者惡意降低流速的演算法。我們的即時演算法憑藉網路流量以及可用頻寬資訊，來判斷使用者是否欺騙。由於近年來可用頻寬測量技術的進步，我們能在短時間內測量到一定準確度的動態可用頻寬。我們選擇 pathChirp 來做我們的測量工具。pathChirp 在我們的實驗中，擁有較高的準確性以及較低的資料量。是比較好的選擇。我們的 irregular throughput detection 演算法在同儕網路上能夠辨認超過90%以上惡意降低流速的使用者並且避免誤判的情形發生。最後，我們的模擬結果證實我們的演算法可以提高multicast系統的效能。

關鍵字

同儕網路；可用頻寬偵測

並列摘要

In recent years, peer-to-peer applications become popular in today’s Internet. However, there is an inherent selfish problem in peer-to-peer networks. Users attempt to strategically interact with other users to gain benefit for themselves. These selfish behaviors will decrease the performance and lead to instability of the network. Therefore, we focus our research on eliminating the effect from selfish users, especially the bandwidth-cheating problem, which means that users decrease throughput intentionally. We aim to design an algorithm to discourage cheating behaviors in peer-to-peer networks. We propose a real-time cheating detection algorithm based on throughput and available bandwidth estimation. Thanks to the rapid development of available bandwidth estimation techniques, we are able to estimate dynamic available bandwidth in short time. We choose pathChirp to estimate available bandwidth based on its better accuracy and less overhead than other estimation tools. Our irregular throughput detection algorithm is able to identify more than 90% cheating peers and avoid misjudging honest users affected by unstable network conditions. Finally, our simulation results show that our algorithm improve the performance in application-layer multicast systems.

並列關鍵字

peer-to-peer netowrk ； available bandwidth estimation

參考文獻

[3] The Kazaa homepage, http://www.kazaa.com

[4] J Strauss, D Katabi, F Kaashoek, ”A measurement study of available bandwidth estimation tools,” in Proceedings of the ACM SIGCOMM, 2003

[6] Vinay J. Ribeiro, Rudolf H. Riedi, Richard G. Baraniuk, Jiri Navratil, Les

[9] A. Downey., “clink: a Tool for Estimating Internet Link Characteristics”, http://rocky.wellesley.edu/downey/clink/, 1999

[11] Jean-Chrysostome Bolot., “End-to-end pakcet depay and loss behavior in the internet,” In Proceedings of ACM SIGCOMM, San Francisco, CA, September 1993

國際替代計量

在同儕網路上以可用頻寬量測機制偵測自私使用者之研究

全文下載

主題瀏覽