Quantum cryptography is a technology becoming mature. Two manufacturers (idQuantique, MagiQ) are offering commercial quantum key distribution systems based on the transmission of weak optical pulses via optical fibres already.

While the basic protocols of quantum key distribution are proven secure there is a great number of cryptographic tasks that have no (known) quantum solution and thus need the help of classical cryptography. The University of Waterloo has a traditional strength in cryptography manifested in its Centre for Applied Cryptographic Research. We thus want to create a continuously operating quantum key distribution link between the Institute for Quantum Computing on the one end and Perimeter Institute on the other. Implementing this link will require input from classical cryptographers and help attract industrial partners interested in quantum technology.

Instead of fibres we chose to implement a free-space optical link based on roof-top telescopes. This choice is motivated by easy installation and reconfigurability and in the view of possible extension and connectivity to satellite based quantum key distribution. Since the atmosphere is not birefringent, polarization is the ideal degree of freedom for a quantum channel with very little decoherence. Even with small telescopes of about 10cm lens diameter one can achieve link transmission of up to 50%.

Entanglement based quantum key distribution does not require fast switching or modulation. A typical compact source for polarization entangled photon pairs using a semiconductor blue laser will be able to deliver a detected pair rate of up to 100 kHz even at moderate link transmission. Future sources based on periodically poled materials may be able to yield even higher rates, in order to get a final key rate that will allow to encrypt all communication between the two locations.
The two observer stations will be based on GPS time references for data collection synchronized to about 20ns. This accuracy should be enough to eliminate the noise caused by atmospheric stray light and detector dark counts. The time-tagged raw data will then go through reconciliation, error correction and privacy amplification, all to be implemented as a protocol stack that is compatible with internet communication.

Having a working quantum key distribution system will allow us to try out new protocols that have not been implemented before. They promise both higher security and immunity against certain deficiencies in the practical realizations of quantum key distributions.

There is a widespread belief that entanglement based quantum key distribution is immune against imperfect sources that emit more than one pair at a time. However, this problem has not been studied in detail and there is no security proofs that take this issue into account. We will study the suitability of decoy techniques to entanglement based systems.

With entanglement based quantum key distribution it is easy to introduce more bases and realize, for example, the six-state protocol, which can give better security thresholds under certain conditions.

The protocols suggested by Boileau et al. [PRL 92, 17901 (2004) are robust against collective noise and, in part, photon loss. With a working quantum key distribution system we can try to implement these protocols.

Current quantum key distribution protocols involve some tacit assumptions, for example that quantum physics is fundamentally correct. Trying to reduce quantum key distribution to fewer assumptions Kent et al. [quant-ph/0405101] have suggested a protocol that only relies on the eavesdroppers inability to signal faster than light. On the working system we may be able to implement this protocol.

Fingerprinting is about testing for equality of bit strings. Quantum fingerprinting tests for equality of qubits. R. Horn has suggested to use the equality test for a new symmetric key distribution protocol.