Anton Trushechkin

Networks of nodes connected by pairwise quantum key distribution (QKD) links are actively developing now. We consider the following problem: Given pairwise secret keys from QKD, how to agree on a common (conference) key for the whole network using classical communication? We propose an algorithm based on spanning tree packing from the graph theory and prove its optimality. The same algorithm can be applied for the GHZ distillation in pair-entangled networks.

We investigate the problem of conference key agreement in pair entangled networks (PEN) where the parties can share bipartite entangled states. In such networks, a source whose global state factorizes into bipartite “pair-entangled network” (PEN) states is distributed to honest parties which can perform local operations and public classical post-processing (LOSR+PP) to establish a shared secret key among each other. In this setting, we derive several new upper bounds on the achievable conference key rate. In particular for pure PEN states we show that the optimal key rate can be achieved by a bipartite QKD strategy.

We investigate the security of a quantum key distribution scheme, where Bob right after each detection announces publicly his choice of measurement basis, and the measurement results if the measurement is performed in the testing basis (used for parameter estimation). Such a scheme saves memory and communication time on both sides by not sending all the classical information only at the end of each block but immediately after each detection. We prove its security and show that this method will not reduce the key rate compared to conventional sifting.

One of the most important requirements for the correct operation of the BB84 protocol is the preparation of the true random state. Most realizations follow this logic: Alice prepares random quantum states, measures them to extract random numbers, and then uses them to modulate the state of the transmitted light. The alternative approach is passive state preparation. It was proposed in 2010 and recently studied for security aspects. The idea is to use the natural phase randomness of the laser pulses to prepare random states. This approach can help to solve the security problem of correlating the state modulation voltage. Originally, the focus was on preparing the polarization state. This required two lasers or an additional intensity modulator. In this work, we use a laser that generates random phase pairs of subsequent pulses as a ready-to-use qubit. This allows us to simplify the Alice device. To perform a full phase characterization, we split a portion of the signal, convert it to polarization, and perform polarization tomography where we postselect four BB84 states. Without a decoy state, this QKD system is well suited for the last mile of a star quantum network with a loss budget of up to 10 dB. We have experimentally demonstrated passive state QKD over 10km deployed and spool fiber obtaining 10-100 bps of secret key correspondingly.

The detection-efficiency mismatch is a common problem in practical quantum key distribution (QKD) systems. The security of quantum key distribution in this case is proved only under the assumption that either the output of the sender side or the input to the receiver side are single-photon signals, which impose a restriction over the class of possible eavesdropping strategies. Here we present a security proof without such a restriction and, thus, solve this important problem and prove the security of quantum key distribution with detection-efficiency mismatch against general attacks. In particular, we propose an adaptation of the decoy state method to the case of detection-efficiency mismatch.