[Freeswitch-users] Encrypted RFC2833 DTMF

[Steve Underwood]

On 06/03/2013 08:19 AM, James Cloos wrote:
> Goog found this:
>
> http://enterprise.huawei.com/ilink/enenterprise/download/HW_U_149094
>
> which says:
>
> ,----< excerpt from HW_U_149094.pdf §35.3 >
> | Huawei NGN Cipher Version1 (HNC1) is a patented algorithm of Huawei
> | and supports the 122/256-bit key.  In encryption, the softswitch
> | (SoftX3000) of Huawei is required to cooperate.  Currently, HNC1 is
> | used in the application scenarios in which the UA5000 works with the
> | SoftX3000.
> `----
>
> It also says that those two products encrypt the rfc2833 traffic with
> that algorithm, w/o implying that it is part of 2833.
>
> As for why it is useful to encrypt the 2833 w/o also encrypting the
> voice, I can only speculate.  Perhaps some idiosyncrasy of PRC law?
> Or to provide better integrity?
>
> -JimC
There is an excellent reason for encrypting DTMF, when encrypting the 
voice is unimportant. If you can compromise a gateway and pick out all 
the DTMF, you might have the ability to recover a lot of passwords used 
for things like phone banking. On the PSTN this kind of thing only 
really works with a focussed hardware attack on the lines to the banking 
system, but with VoIP any compromised node could be a problem.

Steve