Spot Me if You Can: Uncovering Spoken Phrases in Encrypted VoIP Conversations

摘要: Despite the rapid adoption of Voice over IP (VoIP), its security implications are not yet fully understood. Since VoIP calls may traverse untrusted networks, packets should be encrypted to ensure confidentiality. However, we show that when audio is encoded using variable bit rate codecs, lengths can used identify phrases spoken within a call. Our results indicate passive observer from standard speech corpus with an average accuracy 50%, and greater than 90% for some phrases. Clearly, such attack into question efficacy current encryption standards. In addition, examine impact various features underlying on our performance discuss methods mitigation.