Cryptography Reference
In-Depth Information
To characterize realistically the quality of our source without completely re-
constructing the density matrix, we make an assumption on the noise present
in our produced output state. Assuming random (white) noise, our produced
state becomes
+
(
1
−
v)
4
=
v
|
ψ
−
ψ
−
|
ρ
I
.
(3.3)
12
12
12
In this model, the quality of our source is entirely described by the two-
photon visibility
and the pair production rate per second. The overall num-
ber of detected photon pairs was approximately 25,000 pairs per second, and
the average visibility was better than
v
v
=
0
.
95.
3.2.5.2.2 QKD Electronics.
The prototype of the dedicated quantum
key distribution (QKD) hardware currently under development consists of
three main computational components: acquisition of the raw key, genera-
tion of synchronization pulses, and QKD protocol tasks. All three units are
situated on a single printed circuit board. The developed detection logic is
implemented in a FPGA and runs at a sampling frequency of 800 MHz, while
employing a time window of 10 ns for matching the detection events and
synchronization signals.
The board handles the synchronization channel and generates a strong
laser pulse whenever a photon counting event is detected at Alice's site. This
is ensured by a logical OR connection of the detector channels. The synchro-
nization laser pulse at the wavelength of 1550 nm was sent over a separate
single-mode fiber.
A full scale QKD protocol was implemented very recently including data
acquisition, error estimation, error correction, implementing the algorithm
CASCADE [74], privacy amplification, and a protocol authentication algo-
rithm that ensures the integrity of the quantum channel by using a T oplitz
matrix approach. Furthermore, the encryption library modules applied in-
clude one-time pad and AES encryption schemes, the latter allowing key
exchange on a scale determined by the user.
3.2.5.2.3 Results.
The average total quantum bit error rate (QBER) of
the raw key was found to be less than 8% for more than the entire run time
of the experiment. An analysis of the different contributions to the QBER
showed that about 2.6% originate in imperfections of the detection modules
and 1.2% are due to reduced visibility of the entangled state. The rest of the
QBER was attributed to the error produced by the quantum channel. The
average raw key bit rate in our system was found to be about 80 bits/s after
error correction and privacy amplification. This value is mainly limited by
the attenuation on the quantum channel, by the detection efficiency of the
avalanche photodiodes, and by the electronics.
To conclude, polarization-entangled photon systems provide an excellent
alternative for systems based on weak coherent pulses. Our results suggest
that the development of a commercial entanglement based quantum cryp-
tography system is not far away.
