Researchers from the University of z. Baltimore, Maryland, USA, have discovered a vulnerability in the new algorithms for Internet telephony, ensuring high voice quality due to a variable degree of audio compression.
It turned out that the length of the packet, in which different sounds are compressed, you can define whole words and phrases.
Most modern networks, IP-telephony until such methods are protected from interception of conversations, but in the long term, many operators of this type of communication is going to implement new algorithms. Sound compression technology with variable parameters (variable bitrate compression) provides high quality audio at the same link capacity due to the fact that different sounds are encoded packets of varying length. Complex and long combination sound like ...
Researchers were able to demonstrate that the captured packets can be used to decrypt the contents of a conversation. With the help of a special dictionary, compiled on the basis of recorded real negotiations in one of the IP-telephony networks, programmers were able to achieve recognition of words with an accuracy of more than 50%. According to the authors of the work of a group of Johns Hopkins (John Hopkins), is most easily recognized by negotiating with a large number of specialized long term - such terms have a characteristic structure that is easy to identify in the packet stream. It is much harder to identify words and phrases in informal conversations, which uses words from everyday use, often very similar to each other.
Post about the vulnerability of voice compression algorithms, said Phil Zimmerman (Philip Zimmermann), founder of Secure Network Zfone VoIP encryption algorithm and the author of PGP. According to him, trained voice compression algorithms do not look secure. Instead, you should use a partition of compressed voice into packets of equal length, although it may reduce the effectiveness of compression.
Article by John Hopkins and his colleagues about the vulnerability of adaptive algorithms for voice compression in the IP-telephony networks was presented at the Symposium on Security and Privacy 2008 IEEE Symposium on Security and Privacy, held in May this year.
Source:. Security Lab.
No comments:
Post a Comment