Practical Security Analysis Of Quantum Key Distribution System

Information Security is one the most important issue in the study of modern com-munication. Various aspects such as data confidentiality, data integrity and authenticity in a modern communication system are guaranteed by cryptography. For a very long time, classical cryptosystems such as RSA, DES and AES provide information securi-ty depending on the computational complexity of their cryptographic algorithms. Al-though the computer technology made remarkable progress during the past decades, it is not easy to crack a cryptosystem in a reasonable short time. However, since the combination of quantum computer and quantum algorithm can rapidly reduce the diffi-culty of cracking a modern cryptosystem, the fast development of quantum computation technology poses a real challenge to the communication security.The only way to against this super computational power is using the one-time pad cryptosystem, the security of which is unconditional and irrelevant to the computational complexity. The key problem is-as one-time pad requires the users to share and keep updating common secret keys in time-to securely deliver the keys through a public channel, no traditional method can ensure the received keys have not been tapped. In1984, people found the key to solve this problem when Bennett and Brassard proposed the concept of quantum key distribution. After30years’study, the unconditional se-curity of quantum key distribution protocol has been strictly proved, and breakthrough progresses have also been made in the experiments and implementations.Previous study of quantum key distribution security concentrates on the protocols, many of the problems exist in the practical quantum key distribution systems, such as multi-photon, the linear mode of APD single photon detector, etc., are neglected. The study of practical security of quantum key distribution attracts more and more attentions along with the extensive use of quantum cryptography technology today.This paper arranges my study about security of practical quantum key distribution system during my Ph.D programme, the major results are outlined as follows:(1) Modulation errors exist in all kinds of optical modulators. While these com-ponents are placed inside the security zone, these errors could not be exploited by an eavesdropper outside, therefore they reduce the mutual information of the legal users and eavesdropper at the same time. Following this idea, we analyzed how the finite ex-tinction of intensity modulator effects to the practical quantum key distribution systems, and derived the modified secret key rate formula by considering this imperfection.(2) The wavelength dependent property of fiber beam splitter could be exploited by the eavesdropper to fully control Bob’s measurement base choices. We proposed a wavelength attack scheme, which can be applied to all practical quantum key distribu-tion systems with passive state modulation. The proof-of-principle experiment showed that the eavesdropper can obtain all the secret key information at the cost of slightly increasing the quantum bit error rate.(3) Previous studies suggest that light intensity modulation is enough to avoid the potential problems of transmitting the local oscillator along with the signal in a practical continuous variable quantum key distribution system. However, we discovered a new type security loophole exists in heterodyne detection, which can not be fixed by intensity modulation. We the first time proposed an attack scheme aims at the continuous variable quantum key distribution system with heterodyne detection. Our attack is proved to be feasible and renders all of the final keys shared between the legitimate parties without being discovered.(4) The homodyne detection is more widely used than the heterodyne detection in continuous variable quantum key distribution. For preventing the calibration attack, people invented the real-time shot noise modulation method. However, this method is invalid to a stronger wavelength attack based on the wavelength-dependency loophole of the fiber beam splitters. Combined with the experimental data, we prove the feasi-bility and proposed a new defending method which can prevent the calibration attack, saturation attack and wavelength attack at the same time.