Re: Quantum cryptography tackles video

On a sunny day (Wed, 11 May 2005 19:50:04 -0400) it happened Baugh
<baconbaugh@xxxxxxxxxxx> wrote in <NSwge.7744$fY4.2267@xxxxxxxx>:

>Jan Panteltje wrote:
>> On a sunny day (Wed, 11 May 2005 12:31:25 -0400) it happened Baugh
>> <baconbaugh@xxxxxxxxxxx> wrote in <Grqge.3571$sV7.2740@xxxxxxxx>:
> >
>>Baugh:
>>>You are mistaken. The point of quantum crypt. is that should the man
>>>in the middle read the signal he will affect it. You can't anonymously
>>>tap a quantum signal.
>>
>> Wrong, the way it works is that they generate a key, use OTP (one time pad),
>> and xor the data with it.
>> The key is send via the fiber, as polarized photon pairs, so if A measures
>> one polarization, B must have the other.
>> HOWEVER one can buy these units commercially, cut the fiber, decode, fix the
>> data, then re-encode (generate a new pair of photons).
>> The ONLY thing protecting you against that attack is a strict protocol, NOT
>> the quantum stuff.
>> Look it up, sci.crypt, about 2 weeks ago.
>>
>
>Pardon, I didn't read the article. I did attend a couple of seminars on
>the issue. I don't know if this is a case of what I was explained but
>the point of the method is that the pair is entangled.
>
>This means given two sets of polarization measurements, vert vs horiz.
>and l-cir vs r-circ, All corresponding pair mesurements will be
>correlated. However if you intercept one of the pair and measure one
>of the polarization modes (say linear) you cannot use that to replicate
>the that half of an entangled pair. You can reproduce the linear value
>but the circular polarization is now no-longer correlated with that
>of the other of the pair.
>
>The critical point is that what you can measure and then reproduce
>(tapping the line) is mode dependent (circular vs linear polarization,
>or x-spin vs y-spin vs z-spin), but the entanglement correlates all
>modes (or anti-correlates which comes to the same thing). This stronger
>correlation is at the cost of any actual control of which the values
>are. Thus one side doesn't actually send the other a specific signal,
>they both observe the same modes of an entangled pair and get private
>access to the random outcomes.
>
>You could only tap the line if you knew ahead of time which modes the
>two listners would choose to observe. By using a true random number
>generator and the values of previous bits to decide the next they can
>establish a protocol only if you aren't listening to the majority of the
>bits. Their failure is their detection of a tapped line. Their success
>assures that you've not been listening to enough of the signal to get
>the full key. This is a bit simplified but that is the essence of the
>method.
>
>Regards,
>James Baugh
OK, I think I sort of get what you are trying to say.
But this is referred to as 'shared secret'.
In such case you could just as well give both sides a DVD with
'pure random data' and do an OTP.
A lot simpler, and in both cases only the 'shared secret' need to be protected.
Main reason against OTP is key security, now you have the same thing?
.

Relevant Pages