Fast NTRU Encryption in GPU for Secure IoP Communication in Post-quantum Era

Abstract

Internet of People involves a lot of data collection and analysis on human activities, which can pose serious privacy and security issues. Public key cryptography is widely used to securely send secret keys from sensor nodes to cloud server for subsequent decryption of sensors data. However, many popular public key encryption schemes (e.g. RSA and ECC) is deemed insecure in near future, due to algorithms miming on quantum computers. In this paper, we present the implementation of high throughput NTRU encryption scheme in GPU, which can be used to secure the IoP communication against the quantum computer attack On a GTX1080 NVIDIA GPU, our implementation can achieve 508,541 encryptions per second, which is capable in handling massive concurrent secure connections. According to the experimental results, Karatsuba polynomial multiplication proposed in this paper receives 20.6% improvement compared to the flat form implementation. Our implementation is constant time, which provides side channel resistant compared to the other work that utilizes product form and sparse form polynomial multiplication. The proposed solution can be used to secure communication between two parties, which is important for applications like smart home, biometric authentications and body area network