Decrypting Secrets: Deep Learning for Unmasking Malicious Traffic
Shuchi Sethi, Ratna Nitin Patil, Seema Rani, Misbah Anjum, Alka Chaudhary · 2024
The prevailing of encrypted traffic in all network flows, including social media, has significantly enhanced privacy and security for personal data. However, this widespread adoption of encryption poses challenges for the detection of malicious traffic, as adversaries are exploiting this technology to evade detection. Traditional content analysis-based detection techniques are rendered ineffective in the face of encrypted traffic. A survey of current detection methods and technologies underscores the challenges posed by encrypted traffic for existing approaches. This underscores the imperative for sophisticated detection techniques to effectively handle malicious activities within encrypted traffic. In response to this challenge, this paper introduces a framework based on deep learning for the identification of malicious data traffic within the encrypted deep web. The framework’s performance is evaluated using a combination of datasets from five different public sources, and the results are compared with existing detection technologies. The pre-processing allows extracting important features and an application of deep learning techniques give high accuracy.