Efficient anonymous web browsing preventing traffic analysis attacks

Abstract

Anonymous web browsing is an emerging research topic that allows anyone to surf the web privately and securely. However, research on low latency anonymity application including web browsing has received greater attention in order to preserve anonymity but it is quite impractical due to the enormous delay and extra bandwidth demand caused by the current dummy packet padding strategy against traffic analysis attacks. To overcome this problem, we propose a new packet padding strategy to replace the current dominant dummy packet padding strategy for anonymous web browsing systems. The proposed method uses prefetched pages as the cover traffic rather than the dummy packets. With the proposed model, we compared our proposal and the previous scheme in terms of the metric Anonymization Effectiveness (AE) theoretically. Moreover, we defined a metric to measure the efficiency of different padding strategies and established a mathematical model for the traffic analysis attack. Finally, the experimental results have shown that the proposed method is efficient in reducing the delay and bandwidth waste compared to that of dummy packet padding method which confirmed the potentials of the proposed strategy.