Unconditionally secure communication, which has been pursued for thousands of years hasn't, however, yet been reached due to a continuous competition between encryption and hacking. Quantum key distribution (QKD), harnessing the quantum mechanical nature of superposition and noncloning, may promise unconditional security by incorporating the one-time pad algorithm rigorously proved by Claude Shannon. Massive efforts have been made in building practical and commercial QKD systems; in particular, decoy states have been employed to detect photon-number splitting attacks against a single-photon source loophole, and measurement-device-independent (MDI) QKD has further closed all loopholes on the detection side, which leads to a seemingly real-life application. Here, we propose and experimentally demonstrate a MDI-QKD hacking strategy on the trusted-source assumption by using an injection-locking technique. An eaves-dropper (Eve) injects near off-resonance photons with randomly chosen polarization into a sender's laser, where injection locking in a shifted frequency can happen only when Eve's choice matches the sender's state. By setting a shifted window and switching the frequency of photons back afterwards, Eve in principle can obtain all the keys without terminating the real-time QKD. We observe the dynamics of a semiconductor laser with injected photons, and obtain a hacking success rate reaching $60.0\mathrm{%}$ of raw keys. Our results suggest that the spear-and-shield competition of unconditional security may continue until all potential loopholes are discovered and closed ultimately.
Unconditionally secure communication, which has been pursued for thousands of years hasn't, however, yet been reached due to a continuous competition between encryption and hacking. Quantum key distribution (QKD), harnessing the quantum mechanical nature of superposition and noncloning, may promise unconditional security by incorporating the one-time pad algorithm rigorously proved by Claude Shannon. Massive efforts have been made in building practical and commercial QKD systems; in particular, decoy states have been employed to detect photon-number splitting attacks against a single-photon source loophole, and measurement-device-independent (MDI) QKD has further closed all loopholes on the detection side, which leads to a seemingly real-life application. Here, we propose and experimentally demonstrate a MDI-QKD hacking strategy on the trusted-source assumption by using an injection-locking technique. An eaves-dropper (Eve) injects near off-resonance photons with randomly chosen polarization into a sender's laser, where injection locking in a shifted frequency can happen only when Eve's choice matches the sender's state. By setting a shifted window and switching the frequency of photons back afterwards, Eve in principle can obtain all the keys without terminating the real-time QKD. We observe the dynamics of a semiconductor laser with injected photons, and obtain a hacking success rate reaching 60.0\mathrm{%} of raw keys. Our results suggest that the spear-and-shield competition of unconditional security may continue until all potential loopholes are discovered and closed ultimately.
Decoy-state measurement-device-independent quantum key distribution (MDI-QKD) is regarded as one of the most promising protocols since it is physically immune to all fatal detection-side attacks. However, the loopholes in the source side still open side channels to eavesdroppers. Previous work has indicated that a system is insecure if its decoy-state modulation is inaccurate, but a practical method for Eve to exploit the inaccurate modulation is still absent. In this work, we propose a practical hacking method and provide a detailed theoretical analysis. Using our method, we experimentally hacked a MDI-QKD system and successfully obtained all final keys. By demonstrating the hacking experiment, we hope to bring attention to source-side security and provide inspiration for the development of protocols with better source-side security.
Cryptography plays a key role in today's digital society. The most widely used encryption method is the public-key algorithm, but its security relies on unproven computational assumptions. In contrast, quantum key distribution (QKD) is an unbreakable encryption algorithm. QKD has been proven to provide information-theoretic security rigorously based on the laws of physics. Despite the beauty of QKD theory, in practice, QKD still suffers from practical side-channel attacks, due to the imperfections of real-life implementations. Such discrepancy between the theory and the practice is fatal to the security of a QKD system. This thesis addresses a number of the discrepancies in a standard QKD implementation, and it offers efficient countermeasures to resolve these discrepancies. A novel protocol -- measurement-device-independent (MDI) QKD - was proposed to remove all detector side-channel attacks. To apply MDI-QKD in real life, I have proposed practical protocols, conducted the first rigorous finite-key security proof, designed methodologies to optimize parameters, and demonstrated MDI-QKD experimentally. These results allow MDI-QKD mature enough to be implemented in real-life. Based on the framework of MDI-QKD, the only security issue left is the source. To solve the loophole due to the imperfect encoder, I have experimentally implemented a recent protocol, which enables QKD loss-tolerant to source flaws. This experiment for the first time shows secure QKD with imperfect state preparations at long distances. Another vulnerable part in the source is the ability to generate true randomness. I have developed a compact and cost-effective quantum random number generator (QRNG) prototype, which has a real-time generation rate of 1 Gbit/s with excellent stability. The simplicity and speed of my prototype prove the feasibility of a robust, low-cost, and fast QRNG. In summary, MDI-QKD enables new scientific developments in the field of quantum cryptography. This thesis solves the major practical issues in the implementation of MDI-QKD. The combination of MDI-QKD, imperfect sources, and QRNG consists of the foundation of a future practical side-channel-free QKD protocol. The success of this protocol can substitute current classical cryptosystems by offering people an unconditionally secure infrastructure, which will be the ultimate dream for researchers in secure communication.
Quantum key distribution is proven to offer unconditional security in communication between two remote users with ideal source and detection. Unfortunately, ideal devices never exist in practice and device imperfections have become the targets of various attacks. By developing up-conversion single-photon detectors with high efficiency and low noise, we faithfully demonstrate the measurement-device-independent quantum-key-distribution protocol, which is immune to all hacking strategies on detection. Meanwhile, we employ the decoy-state method to defend attacks on a nonideal source. By assuming a trusted source scenario, our practical system, which generates more than a 25 kbit secure key over a 50 km fiber link, serves as a stepping stone in the quest for unconditionally secure communications with realistic devices.
Quantum key distribution (QKD) enables two legitimate parties to share a secret key even in the presence of an eavesdropper. The unconditional security of QKD is based on the fundamental laws of quantum physics. Original security proofs of QKD are based on a few assumptions, e.g., perfect single photon sources and perfect single-photon detectors. However, practical implementations of QKD systems do not fully comply with such assumptions due to technical limitations. The gap between theory and implementations leads to security loopholes in most QKD systems, and several attacks have been launched on sophisticated QKD systems. Particularly, the detectors have been found to be the most vulnerable part of QKD. Much effort has been put to build side-channel-free QKD systems. Solutions such as security patches and device-independent QKD have been proposed. However, the former are normally ad-hoc, and cannot close unidentified loopholes. The latter, while having the advantages of removing all assumptions on devices, is impractical to implement today. Measurement-device-independent QKD (MDI-QKD) turns out to be a promising solution to the security problem of QKD. In MDI-QKD, all security loopholes, including those yet-to-be discovered, have been removed from the detectors, the most critical part in QKD. In this thesis, we investigate issues related to the practical implementation and security of MDI-QKD. We first present a demonstration of polarization-encoding MDI-QKD. Taking finite key effect into account, we achieve a secret key rate of 0.005 bit per second (bps) over 10 km spooled telecom fiber, and a 1600-bit key is distributed. This work, together with other demonstrations, shows the practicality of MDI-QKD. Next we investigate a critical assumption of MDI-QKD: perfect state preparation. We apply the loss-tolerant QKD protocol and adapt it to MDI-QKD to quantify information leakage due to imperfect state preparation. We then present an experimental demonstration of MDI-QKD over 10 km and 40 km of spooled fiber, which for the first time considers the impact of inaccurate polarization state preparation on the secret key rate. This would not have been possible under previous security proofs, given the same amount of state preparation flaws.
Continuous-variable measurement-device-independent quantum key distribution (CV-MDI QKD) can defend all detection-side attacks effectively. Therefore, the source side is the final battlefield for performing quantum hacking attacks. This paper investigates the practical security of a CV-MDI QKD system under a light-injection attack. Here, we first describe two different light-injection attacks, i.e., the induced-photorefractive attack and the strong-power injection attack. Then, we consider three attack cases where Eve only attacks one of the parties or both parties of the CV-MDI QKD system. Based on the analysis of the parameter estimation, we find that the legitimate communication parties will overestimate the secret key rate of the system under the effect of a light-injection attack. This opens a security loophole for Eve to successfully obtain secret key information in a practical CV-MDI QKD system. In particular, compared to the laser-damage attack, the above attacks use a lower power of injected light and have a more serious effect on the security of the system. To eliminate the above effects, we can enhance the practical security of the system by doping the lithium niobate material with various impurities or by using protective devices, such as optical isolators, circulators, optical power limiters, and narrow-band filters. Apart from these, we can also use an intensity monitor or a photodetector to detect the light-injection attack.
Measurement-device-independent quantum key distribution (MDI-QKD) can be immune to all detector side-channel attacks and guarantee the information-theoretical security even with uncharacterized single photon detectors. MDI-QKD has been demonstrated in both laboratories and field-tests by using attenuated lasers combined with the decoy-state technique. However, it is a critical assumption that the sources used by legitimate participants are trusted in MDI-QKD. Hence, it is possible that a potential security risk exists. Here we propose a new scheme of polarization-encoding-based MDI-QKD with a single untrusted source, by which the complexity of the synchronization system can be reduced and the success rate of the Bell-state measurement can be improved. Meanwhile, the decoy-state method is employed to avoid the security issues introduced by a non-ideal single photon source. We also derive a security analysis of the proposed system. In addition, it seems to be a promising candidate for the implementation for QKD network in the near future.