Objective To address the limitation imposed by a high-speed, high-resolution digital-to-analog converter (DAC) on the quantum noise randomized cypher (QNRC) research, a novel scheme using the structure of paralleled intensity modulators is proposed, increasing the number of ciphertext states considerably.
Methods A detailed explanation of the intensity shift keying (ISK)-QNRC encryption principle demonstrates that an ISK-QNRC with paralleled modulators is feasible. In this study, we established a system which employs optical domain decryption and direct detection. On the receiver’s side, a running key is modulated on a local oscillator as the decryption signal. After matching the power of the decryption signal and ciphertext signal, which is from the sender’s side, these two signal can be converted into a binary electrical signal via a balanced photonics detector. The plaintext is obtained after XOR for the binary electrical signal with the least significant bit of the running key.
Results and Discussions We establish an optical communication system (Fig.5) and configure it properly (Table.1) using professional simulation software VPItransmission Maker Optical System 9.1. Simulation results (Fig.6) show that the ISK-QNRC system with the mesoscopic power of -20 dBm, a transmission distance of 500 km, the bit rate of 10 Gbit/s, and the number of ciphertext states of 2 10-1 can be realized, while a plaintext cannot be obtained using the wire-tapper without the running key. We discuss the effect of the received power, mesoscopic power, and transmission length on the error performance. As is shown in Fig.7, the bit error rate (BER) decreases as the received power increases at a specific transmission length (B2B, 500 km, 1000 km) and a specific mesoscopic power (-20 dBm). In addition, when the BER is 10 -10, the power penalty for 500 km is approximately 1 dBm compared with B2B and that for 1000 km is approximately 1.35 dBm compared with 500 km. Fig.8 describes the relationship between the BER and the mesoscopic power at a specific transmission length (B2B, 500 km) and a specific received power (-10 dBm, -15 dBm). When the transmission length is 500 km, the error performance improves with increasing received power, and when the received power is -15 dBm, the error performance improves with decreasing transmission length. Although the transmission length is longer when the mesoscopic power is greater than -18 dBm, the received power of the -10-dBm signal has a lower BER than that of B2B signal. Here, the received power plays a dominant role in improving the error performance. By contrast, when the mesoscopic power is less than -18 dBm, the transmission length plays a dominant role in improving the error performance. Fig.9 describes the relationship between the transmission length and BER at specific mesoscopic power (-15 dBm, -20 dBm) and specific received power (-10 dBm, -15 dBm). It is well known that the BER always increases as the transmission length increases because a longer distance means more amplified spontaneous emission noise caused by erbium-doped fiber amplifier. By setting different variable optical attenuator parameters at the receiver’s side, we find that the BER is sensitive to the power difference between ciphertext and decryption signals.
Conclusions This study aims to overcome the limitation of high-speed and high-resolution DACs. Therefore, we propose the parallel intensity modulation, optical domain decryption, and direct detection methods to realize the parallel ISK-QNRC scheme. The effect of changes in three important physical parameters (received power, mesoscopic power, and transmission length) on error performance is discussed and analyzed, and the results of three performance indicators of effectiveness, reliability, and security are obtained (mesoscopic power of -20 dBm; transmission length of 500 km; bit rate of 10 Gbit/s; number of ciphertext states of 2 10-1, BER of 10 -10). In addition, we observe that the error performance of this scheme is sensitive to the power difference between the ciphertext and decryption signals. In an actual operation, this may affect the error performance of the entire system, but this effect can be eliminated by precisely matching the power of the two signals.
