VoIP networks use the real time protocol (RTP) to send voice over the Internet. Although RTP chunks, encodes and packetizes audio in a simple but efficient way, it has no consideration for security and data integrity.
In fact the lack of security makes RTP (and VoIP in general) a target of call “wiretapping”. How is this possible, you may ask? Well folks, read this below to try it yourself!
- Download and Install wireshark (http://www.wireshark.org)
- Run Wireshark and start capturing packets with the appropriate NIC
- Place a G.711 mu law call using Kapanga
- On Wireshark, stop capturing packets
- Run statistics/RTP/Show all streams
- Click on the streams in the list, do a stream analysis and save the payload as an audio file
- Congrats: you have a recording of the call (one stream per direction)
How can we prevent this from happening? SRTP is an upgrade to RTP that provides security through encryption and authentication. Encryption specifically guarantees that the audio is unintelligible by the time it is sent out. So if we go back to item #3 on the list above we place an SRTP call instead, the recorded audio will sound like “noise”, this is because SRTP rearranges information in the payload using the AES algorithm.
So how’s your softphone’s SRTP support? As it turns out we spent quite a bit of time supporting this feature. And yes, it is available in the public site. Enjoy!