In this project, Nomura HD, Nomura Securities, NICT, Toshiba and NEC have jointly verified the applicability of
quantum cryptographic communication to the financial sector, which is the only cryptographic communication method guaranteed to be undecipherable by any third party (eavesdropper who has unbounded computational power, including large scale quantum computing).
Figure 1 shows the outline of the system that used the joint verification system. We verified low latency and large-volume tolerance utilizing cryptographic equipment using shared random numbers (keys) from
Quantum Key Distribution (QKD) equipment that shares keys by placing random number information on photons.
For this purpose, Nomura HD and Nomura Securities developed an application that simulates financial transactions between investors and securities companies on the
Tokyo QKD Network, the test communications network environment in which NICT introduced and built a QKD device in 2010. The application generates simulated data in accordance with the FIX protocol, which is the standard format used for actual stock orders.
In addition, NICT has been examining the data-encryption method combined with QKD for social implementation, and has adopted two encryption methods: the
one time pad (OTP) method and the Advanced Encryption Standard (AES) method.
OTP is a highly secure encryption method (information theoretic security) in which the encryption cannot be decrypted by a third party with infinite computing power. However, since it requires the same amount of encryption keys as those used in transmission, it tends to consume a large amount of keys, and as a result, there is a risk of key exhaustion. This time, AES was used in conjunction with OTP in order to prepare for key depletion. In addition, a
high-speed OTP device newly developed by NICT to realize a high throughput of Gbps was adopted for the test.
Unlike OTP, AES is not information-theoretically secure. It is a secure encryption method that depends on the computational complexity that requires astronomical calculations to decrypt the data. In this use case, we thought that the AES method would have sufficient security strength by updating the common key generated by QKD in a short period of time. Therefore, we chose AES (AES256), which uses a key length of 256 bits, as an alternative to OTP. We adopted two types of AES256 implementations: A software-based implementation (
SW-AES) and a low-latency
network encryptor (COMCIPHER-Q) developed by NEC. We measured the communication performance of the three types of encryption methods, high-speed OTP, SW-AES, and COMCIPHER-Q.
Based on the keys exchanged by the high-speed QKD equipment developed by Toshiba and the QKD equipment developed by NEC, we tested cases in line with actual stock trading operations and measured the response times of several different data encryption methods during large-volume data transmission. In this way, we verified the practicality of QKD systems and each encryption method. Specifically, we examined the effects of the three types of encryption methods, high-speed OTP, SW-AES, and COMCIPHER-Q, on the total data capacity of the FIX messages transmitted in the stock trading operations of securities companies per day and the response times measured assuming a data transmission volume several tens of times larger than the total data capacity. By comparing and verifying the practicality of QKD systems and data encryption methods during large-volume data transmission in line with these specific test cases for stock trading, we have obtained important suggestions for the future social implementation of quantum cryptography in various fields, including non-financial fields.
