At 03:17 PM 6/26/00 BST, Mo Rainey wrote:
>Does anyone know where I can get an encryption chip that I can build into a
>radio system for secure voice communications?
>
>I need something simple ( i.e. data + key > in, encrypted data > out). I
>plan to put the audio input through an ADC (Analogue to Diigital Converter),
>so it will be a digital data stream going into the chip .I need it to use
>secret key technology so that both radio sets have the same key for
>encrypting and decrypting. It doesn't have to be the most secure thing in
>the world anything that would take more than a week to crack would do,
>sacrifice the security for the speed keeping the cost down
Finding a chip to do the encryption is the least of your problems -
you need to figure out the system design first, and that means you need
to be clear about what you're trying to accomplish with your system.
Have you figured out what you're doing to modulate the digital signal onto
the radio?
What bit rate do you expect to use? Are you using a dedicated chip for the
voice compression (or are you just using lots of bandwidth instead of
compressing?)
Encryption is usually much less work than good voice compression;
a general purpose CPU or a DSP both have more than enough horsepower,
so if you're doing voice compression in one of them, you can do the encryption
in software rather than adding a dedicated power-consuming chip.
If you're saying it needs to be secret key, with the same key in both
directions,
but doesn't need to be very secure, then you still need to think about
your requirements. Encryption that's strong doesn't take any more work than
wimpy encryption, particularly for algorithms like RC4 - as long as you use
it correctly,
a 128-bit key takes no more work than a 40-bit key - but you can't use the
same key in both directions. On the other hand, user-selected passwords
don't take a week to decrypt - they're usually either wimpy enough to take
only a few seconds/minutes to guess, or strong enough to take a very long
time.
You're much better off using Diffie-Hellman key exchange, so nobody can
decrypt your
conversation, and optionally using a secret-key passphrase for
authentication.
(In a radio environment, it's very hard to do man-in-the-middle attacks,
so the main threat is somebody impersonating the other speakers;
you might or might not need authentication depending on whether you
expect to recognize their voice or not.)
Thanks!
Bill
Bill Stewart, [EMAIL PROTECTED]
PGP Fingerprint D454 E202 CBC8 40BF 3C85 B884 0ABE 4639
