[VOIPSEC] Performance/QoS implications of Secure SIP?

[Dan Wing]

> As far as securing the voice itself, there is a new up and coming
> protocol called Secure-RTP or SRTP which allows encryption of RTP
> packets.  The QoS implications of this are as of yet unknown since
> it is not in widespread deployment in the industry.  You can follow
> this mailing list closely because there is a lot of discussion on
> SRTP here which you might find educaitonal.  Good luck!

Tao, see:
http://sourceforge.net/mailarchive/forum.php?thread_id=5731549&forum_id=9697
I haven't looked at the analysis, but it's a datapoint you may want to
consider.

Also consider that if the same processor is doing your voice compression
(codec) and your SRTP encryption/authentication, the MIPS consumed
doing compression are MIPS saved when doing AES-128 and SHA-1 (because
the payload is smaller).  Likewise, if you're not doing much compression
for a call, the MIPS saved by doing light compression are expended when
doing AES-128 and SHA-1 (because the payload is larger).

-d