Pre-shared keys (PSK) have been widely used in network security. Nonetheless, existing PSK solutions are not scalable. Moreover, whenever a new user joins a network, PSK requires an existing user to get a new key before they are able to communicate with the new user. The key issue is how to distribute the PSK between different users. Here, we solve this problem by proposing a new protocol called Distributed Symmetric Key Establishment (DSKE)1. DSKE has the advantage of being scalable. Unlike standard public key infrastructure (PKI) which relies on computational assumptions, DSKE provides information-theoretic security in a universally composable security framework. Specifically, we prove the security (correctness and confidentiality) and robustness of this protocol against a computationally unbounded adversary, who additionally may have fully compromised a bounded number of the intermediaries and can eavesdrop on all communication. DSKE also achieves distributed trust through secret sharing.We present several implementations of DSKE in real environments, such as providing client services to link encryptors, network encryptors, and mobile phones, as well as the implementation of intermediaries, called Security Hubs, and associated test data as evidence for its versatility. As DSKE is highly scalable in a network setting with no distance limit, it is expected to be a cost-effective quantum-safe cryptographic solution to the network security threat presented by quantum computers.

We present a design for a quantum key distribution (QKD) system in a Sagnac loop configuration, employing a novel phase modulation scheme based on frequency shift, and demonstrate stable BB84 QKD operation with high interference visibility and low quantum bit error rate (QBER). The phase modulation is achieved by sending two light pulses with a fixed time delay (or a fixed optical path delay) through a frequency shift element and by modulating the amount of frequency shift. The relative phase between two light pulses upon leaving the frequency-shift element is determined by both the time delay (or the optical path delay) and the frequency shift, and can therefore be controlled by varying the amount of frequency shift. To demonstrate its operation, we used an acousto-optic modulator (AOM) as the frequency-shift element, and vary the driving frequency of the AOM to encode phase information. The interference visibility for a 40 km and a 10 km fiber loop is 96% and 99%, respectively, at single photon level. We ran BB84 protocol in a 40-km Sagnac loop setup continuously for one hour and the measured QBER remained within the 2%-5% range. A further advantage of our scheme is that both phase and amplitude modulation can be achieved simultaneously by frequency and amplitude modulation of the AOM's driving signal, allowing our QKD system the capability of implementing other protocols, such as the decoy-state QKD and the continuous-variable QKD. We also briefly discuss a new type of Eavesdropping strategy ("phase-remapping" attack) in bidirectional QKD system

Decoy state quantum key distribution (QKD) has been proposed as a novel approach to improve dramatically both the security and the performance of practical QKD set-ups. Recently, many theoretical efforts have been made on this topic and have theoretically predicted the high performance of decoy method. However, the gap between theory and experiment remains open. In this paper, we report the first experiments on decoy state QKD, thus bridging the gap. Two protocols of decoy state QKD are implemented: one-decoy protocol over 15 km of standard telecom fiber, and weak+vacuum protocol over 60 km of standard telecom fiber. We implemented the decoy state method on a modified commercial QKD system. The modification we made is simply adding commercial acousto-optic modulator (AOM) on the QKD system. The AOM is used to modulate the intensity of each signal individually, thus implementing the decoy state method. As an important part of implementation, numerical simulation of our set-up is also performed. The simulation shows that standard security proofs give a zero key generation rate at the distance we perform decoy state QKD (both 15 km and 60 km). Therefore decoy state QKD is necessary for long distance secure communication. Our implementation shows explicitly the power and feasibility of decoy method, and brings it to our real-life

We propose a wide class of distillation schemes for multi-partite entangled states that are CSS-states. Our proposal provides not only superior efficiency, but also new insights on the connection between CSS-states and bipartite graph states. We then consider the applications of our distillation schemes for two cryptographic tasks - namely, (a) conference key agreement and (b) quantum sharing of classical secrets. In particular, we construct "prepare-and-measure" protocols. Also we study the yield of those protocols and the threshold value of the fidelity above which the protocols can function securely. Surprisingly, our protocols function securely even when the initial state does not violate the standard Bell-inequalities for GHZ states. Experimental realization involving only bipartite entanglement is also suggested

We present explicit bounds on the classical communication cost and inefficiency of entanglement dilution via the Lo-Popescu protocol, for the case of two-term (single-qubit) entangled states. By considering a two-stage dilution, we consequently use prior results to obtain meaningful bounds on the classical communication cost and inefficiency of dilution between two-term partially entangled states.

In this presentation, we show some counter-examples to a naive belief that the security of QKD is based on no-cloning theorem. One example is shown by explicitly proving that one can indeed generate an unconditionally secure key from Alice's two-photon emission part in "SARG04 protocol" proposed by V. Scarani et al, in Phys. Rev. Lett. 92, 057901 (2004). This protocol differs from BB84 only in the classical communication. It is, thus, interesting to see how only the classical communication of QKD protocol might qualitatively change its security. We also show that one can generate an unconditionally secure key from the single to the four-photon part in a generalized SARG04 that uses six states. Finally, we also compare the bit error rate threshold of these protocols with the one in BB84 and the original six-state protocol assuming a depolarizing channel

This paper prove the security of the Bennett-Brassard (BB84) quantum key distribution protocol in the case where the source and detector are under the limited control of an adversary. This proof applies when both the source and the detector have small basis-dependent flaws, as is typical in practical implementations of the protocol. The estimation of the key generation rate in some special cases: sources that emit weak coherent states, detectors with basis-dependent efficiency, and misaligned sources and detectors.

Hwang (2003) has proposed a decoy state method in quantum key distribution (QKD). In Hwang's proposal, the average photon number of the decoy state is about two. This paper propose a new decoy state scheme using vacua or very weak coherent states as decoy states and discuss its advantages.

Suppose two distant observers, Alice and Bob, share some form of entanglement - quantum correlations - in some bipartite pure quantum states. They may apply local operations and classical communication to convert one form of entanglement to another. Since entanglement is regarded as a resource in quantum information processing, it is an important question to ask how much classical communication, which is also a resource, is needed in the inter-conversion process of entanglement. In this paper, we address this important question in the many-copy case. The inter-conversion process of entanglement is usually divided into two types: concentrating the entanglement from many partially entangled states into a smaller number of maximally entangled states (i.e., singlets) and the reverse process of diluting singlets into partially entangled states. It is known that entanglement concentration requires no classical communication, but the best prior art result for diluting to N copies of a partially entangled state requires an amount of communication on the order of /spl radic/N. Our main result is to prove that this prior art result is optimal up to a constant factor; any procedure for approximately creating N partially entangled states from singlets requires /spl Omega/(/spl radic/N) bits of classical communication. Previously not even a constant bound was known for approximate entanglement transformations. We also prove a lower bound on the inefficiency of the process: to dilute singlets to N copies of a partially entangled state, the entropy of entanglement must decrease by /spl Omega/(/spl radic/N). Moreover, we introduce two new tools - /spl delta/-significant subspaces and the standard form protocol reduction in entanglement manipulations. We hope that these two new tools will be useful in other work in quantum information theory.

Shor and Preskill (see Phys. Rev. Lett., vol.85, p.441, 2000) have provided a simple proof of security of the standard quantum key distribution scheme by Bennett and Brassard (1984) by demonstrating a connection between key distribution and entanglement purification protocols (EPPs) with one-way communications. Here, we provide proofs of security of standard quantum key distribution schemes, Bennett and Brassard and the six-state scheme, against the most general attack, by using the techniques of two-way entanglement purification. We demonstrate clearly the advantage of classical post-processing with two-way classical communications over classical post-processing with only one-way classical communications in quantum key distribution (QKD). This is done by the explicit construction of a new protocol for (the error correction/detection and privacy amplification of) Bennett and Brassard that can tolerate a bit error rate of up to 18.9%, which is higher than what any Bennett and Brassard scheme with only one-way classical communications can possibly tolerate. Moreover, we demonstrate the advantage of the six-state scheme over Bennett and Brassard by showing that the six-state scheme can strictly tolerate a higher bit error rate than Bennett and Brassard. In particular, our six-state protocol can tolerate a bit error rate of 26.4%, which is higher than the upper bound of 25% bit error rate for any secure Bennett and Brassard protocol. Consequently, our protocols may allow higher key generation rate and remain secure over longer distances than previous protocols. Our investigation suggests that two-way entanglement purification is a useful tool in the study of advantage distillation, error correction, and privacy amplification protocols.

We analyze a generalization of Huffman coding to the quantum case. In particular, we notice various difficulties in using instantaneous codes for quantum communication. Nevertheless, for the storage of quantum information, we have succeeded in constructing a Huffman coding inspired quantum scheme. The number of computational steps in the encoding and decoding processes of N quantum signals can be made to be of polylogarithmic depth by a massively parallel implementation of a quantum gate array. This is to be compared with the O(N/sup 3/) computational steps required in the sequential implementation by Cleve and DiVincenzo (see Phys. Rev., vol.A54, p.2636, 1996) of the well-known quantum noiseless block-coding scheme of Schumacher. We also show that O(N/sup 2/(log N)/sup a/) sequential computational steps are needed for the communication of quantum information using another Huffman coding inspired scheme where the sender must disentangle her encoding device before the receiver can perform any measurements on his signals.