Hybrid approach of parallel Implementation on CPU-GPU for high speed ECDSA verification

Abstract

Since the advent of Deep Belief Network (DBN) deep learning technology in 2006, Artificial Intelligence (AI) technology has been utilized in various convergence areas, such as autonomous driving and medical care. Some services requiring fast decision making and action typically work seamlessly with edge computing service model. As an example of an edge computing service model, in autonomous driving of a connected vehicle with Vehicle-to-Everything (V2X) communication, Road Side Unit (RSU) as an edge computing device should process V2X messages sent by vehicles or other RSUs rapidly. IEEE 1609.2 standard provides application message security technology to ensure the security and reliability of V2X communication messages. It uses Elliptic Curve Digital Signature Algorithm (ECDSA) signatures based on the NIST p256 curve for message authenticity. In this dissertation, considering the expected number of vehicles, message delivery rate, and wireless channel (IEEE 802.11p) capacity of the environment such as the intersection in rush hour, we investigate that RSU should verify 3500 ECDSA signatures per second without the help of hardware based cryptographic accelerator. For the requirement, we propose a hybrid approach of parallel ECDSA verification at high speed using CPU and GPU, simultaneously. In this proposed method, the characteristics of CPU and GPU such as signature verification rate and latency are analyzed, and based on the analysis result, the scheduling algorithm for parallel processing on CPU and GPU is configured to maximize the verification performance. Moreover, we implemented the proposed method in various modern computing environments for RSU and edge computing devices. Through the experiments, we reach the conclusion that GPU can contribute to the required performance of ECDSA signature verification in RSU platform, which could not satisfy the above throughput only with CPU. The target platform with Intel Pentium E6500 CPU and GeForce GTX650 GPU can verify 5667 signatures per second with 30% utilization, while CPU in the platform can process only 2668 signatures. Even in a higher performance edge computing device, we examine experimentally that the performance can be further improved by using the proposed hybrid approach.