With photons being the only available candidates for long-distance quantum communication, most quantum cryptographic devices are physically realized as optical systems that operate a security protocol based on the laws of quantum mechanics. But to finally yield a stream of bits (secret key) usable for encryption, a quantum-to-classical transition is required. Synchronization of electronic & optoelectronic components involved in such tasks thus becomes a necessary and important step. However, it also opens up the possibility of timing-based loopholes and attacks.
With photons being the only available candidates for long-distance quantum communication, most quantum cryptographic devices are physically realized as optical systems that operate a security protocol based on the laws of quantum mechanics. But to finally yield a stream of bits (secret key) usable for encryption, a quantum-to-classical transition is required. Synchronization of electronic & optoelectronic components involved in such tasks thus becomes a necessary and important step. However, it also opens up the possibility of timing-based loopholes and attacks.
The peculiar properties of quantum mechanics enable possibilities not allowed by classical physics. In particular, two parties can generate a random, secret key at a distance, even though an eavesdropper can do anything permitted by the laws of physics. Measuring the quantum properties of the signals generating the key, would ultimately change them, and thus reveal the eavesdropper’s presence. This exchange of a random, secret key is known as quantum cryptography.\n\nQuantum cryptography can be, and has been proven unconditionally secure using perfect devices. However, when quantum cryptography is implemented, one must use components available with current technology. These are usually imperfect. Although the security of quantum cryptography has been proven for components with certain imperfections, the question remains: can quantum cryptography be implemented in a provable, unconditionally secure way, using components available with current technology? This thesis contains both a theoretical, and an experimental contribution to the answer of this question. On the experimental side, components used in, and complete quantum cryptography systems have been carefully examined for security loopholes. In particular, it turned out that two commercial quantum cryptography systems contained loopholes, which would allow an eavesdropper to capture the full secret key, without exposing her presence. Furthermore, this detector control attack could be implemented with current technology. The attack is applicable against a variety of quantum cryptography implementations and protocols.\n\nThe theoretical contribution consists of security proofs for quantum cryptography in a very general setting. Precisely, the security is proven with arbitrary individual imperfections in the source and detectors. These proofs should make it possible to use a wide array of imperfect devices in implementations of quantum cryptography.\n\nFinally, a secure detection scheme is proposed, immune to the detector control attack and compatible with those security proofs. Therefore, if this scheme is implemented correctly, it offers provable security.
The use of quantum mechanical signals in communication opens up the opportunity to build new communication systems that accomplishes tasks that communication with classical signals structures cannot achieve. Prominent examples are Quantum Key Distribution Protocols, which allows the generation of secret keys without computational assumptions of adversaries. Over the past decade, protocols have been developed that achieve tasks that can also be accomplished with classical signals, but the quantum version of the protocol either uses less resources, or leaks less information between the involved parties. The gap between quantum and classical can be exponential in the input size of the problems. Examples are the comparison of data, the scheduling of appointments and others. Until recently, it was thought that these protocols are of mere conceptual value, but that the quantum advantage could not be realized. We changed that by developing quantum optical versions of these abstract protocols that can run with simple laser pulses, beam-splitters and detectors. [1–3] By now the first protocols have been successfully implemented [4], showing that a quantum advantage can be realized. The next step is to find and realize protocols that have a high practical value.
Communications in secrecy are often required in many commercial and military applications. Unfortunately, many cryptographic schemes in use today such as public-key cryptography based on the RSA algorithm would be broken with either unanticipated advances in hardware and algorithm or the advent of quantum computers. Quantum cryptography, on the other hand, has been proven secure even against the most general attack allowed by the laws of physics and is a promising technology poised for widespread adoption in realistic cryptographic applications. Quantum cryptography allows two parties to expand on a secret key that they have previously shared. Various quantum cryptographic protocols have been proposed to perform this task. In this paper, we survey some popular quantum cryptographic protocols (including the famous Bennett-Brassard 1984 protocol) and discuss their security. Specifically, we consider their security in two cases: the ideal case where a perfect single-photon source is used and the practical case where a realistic laser source is used. We compare the protocols and find that the efficient six-state protocol outperforms the others both in the tolerable quantum bit error rate and in the key generation rate when a realistic laser source is used.
Quantum cryptography or quantum key distribution (QKD) applies fundamental laws of quantum physics to guarantee secure communication. The security of quantum cryptography was proven in the last decade. Many security analyses are based on the assumption that QKD system components are idealized. In practice, inevitable device imperfections may compromise security unless these imperfections are well investigated.\n\nA highly attenuated laser pulse which gives a weak coherent state is widely used in QKD experiments. A weak coherent state has multi-photon components, which opens up a security loophole to the sophisticated eavesdropper. With a small adjustment of the hardware, we will prove that the decoy state method can close this loophole and substantially improve the QKD performance. We also propose a few practical decoy state protocols, study statistical fluctuations and perform experimental demonstrations. Moreover, we will apply the methods from entanglement distillation protocols based on two-way classical communication to improve the decoy state QKD performance. Furthermore, we study the decoy state methods for other single photon sources, such as triggering parametric down-conversion (PDC) source. Note that our work, decoy state protocol, has attracted a lot of scientific and media interest. The decoy state QKD becomes a standard technique for prepare-and-measure QKD schemes.\n\nAside from single-photon-based QKD schemes, there is another type of scheme based on entangled photon sources. A PDC source is commonly used as an entangled photon source. We propose a model and post-processing scheme for the entanglement-based QKD with a PDC source. Although the model is proposed to study the entanglement-based QKD, we emphasize that our generic model may also be useful for other non-QKD experiments involving a PDC source. By simulating a real PDC experiment, we show that the entanglement-based QKD can achieve longer maximal secure distance than the single-photon-based QKD schemes.\n\nWe propose a time-shift attack that exploits the efficiency mismatch of two single photon detectors in a QKD system. This eavesdropping strategy can be realized by current technology. We will also discuss counter measures against the attack and study the security of a QKD system with efficiency mismatch detectors.
The security of keys in quantum cryptography is based on fundamental quantum mechanical exclusions (the exclusion of cloning and copying of nonorthogonal quantum states. The physical type of a quantum object that carries information (photon, electron, atom, etc.) is insignificant; only its state vector is important. In relativistic quantum cryptography for open space, both the time of the information carrier (photon that propagates with the extremely allowable velocity in a vacuum) and its quantum state are of fundamental importance. Joint fundamental constraints that are dictated by both special relativity and quantum mechanics on the discrimination of nonorthogonal quantum states allow one to formulate fundamentally new key distribution protocols that are stable against any attacks on a key and guarantee the security of keys for a nonstrictly single-photon source and any losses in the communication channel. Although this protocol is a real-time protocol in the Minkowski space-time, where the attack to the communication channel is detected by the delay of eavesdropper measurement results, the protocol does not require clock synchronization on the transmitter and receiver sides .
Quantum-based cryptographic protocols are often said to enjoy security guaranteed by the fundamental laws of physics. However, even carefully designed quantum-based cryptographic schemes may be susceptible to subtle attacks that are outside the original design. As an example, we give attacks against a recently proposed ``secure communication using mesoscopic coherent states'', which employs mesoscopic states, rather than single-photon states. Our attacks can be used either as a known-plaintext attack or in the case where the plaintext has not been randomized. One of our attacks requires beamsplitters and the replacement of a lossy channel by a lossless one. It is successful provided that the original loss in the channel is so big that Eve can obtain 2k copies of what Bob receives, where k is the length of the seed key pre-shared by Alice and Bob. In addition, substantial improvements over such an exhaustive key search attack can be made, whenever a key is reused. Furthermore, we remark that, under the same assumption of a known or non-random plaintext, Grover's exhaustive key search attack can be applied directly to "secure communication using mesoscopic coherent states", whenever the channel loss is more than 50 percent. Therefore, as far as information-theoretic security is concerned, optically amplified signals necessarily degrade the security of the proposed scheme, when the plaintext is known or non-random. Our attacks apply even if the mesoscopic scheme is used only for key generation with a subsequent use of the key for one-time-pad encryption. Studying those attacks can help us to better define the risk models and parameter spaces in which quantum-based cryptographic schemes can operate securely. Finally, we remark that our attacks do not affect standard protocols such as Bennett-Brassard BB84 protocol or Bennett B92 protocol, which rely on single-photon signals.