Context-aware Multimedia Encryption in Mobile Platforms

Voice over IP (VoIP) is a part of multimedia content and this communication is growing rapidly in portable devices due to its attractive features. Secure communication is critical as most of such communications are routed through public channels. Establishing secure VoIP communication is computationally expensive. Modern high strength encryption algorithms such as AES, RSA, Serpent, etc. need high computational power for encryption. On the mobile platform this is a significant factor due to its constrained power resources. In this work we are presenting a novel idea of encrypting VoIP speech data while conserving resources without compromising the security of the VoIP communication. First, the algorithm detects the content of the speech data and identifies its context. Then sensitive sections of the conversation are separated. In the last step, these sections are encrypted with a high strength cryptographic algorithm while other parts are encrypted with a less strength algorithm for better performance. In this way unnecessary encryption power is spared. This will conserve computational power while ensuring the security of sensitive information. Also, we discuss some of the crucial potential attack vectors and their defense mechanisms. Finally, we implemented a prototype to test our concept and we observed an average of 39% reduction in computational time.&