A Micro-Benchmark of ECDSA Curves and Ed25519 for Low-Latency V2X Message Authentication

Abstract

Vehicle-to-Everything (V2X) communication requires robust security measures to ensure the integrity and authenticity of critical information. Although elliptic curve cryptography (ECC) is widely used to meet these requirements, signature verification latency remains a critical bottleneck, especially in high-traffic environments where processing must be completed within a strict 100-millisecond time window. The existing literature lacks controlled system-level comparisons of modern cryptographic curves. To fill this gap, this research proposes a controlled benchmark method using the Java Microbenchmark Harness (JMH) and Bouncy Castle provider to evaluate performance differences between the Elliptic Curve Digital Signature Algorithm (ECDSA) (secp256r1 and brainpoolP256r1) and the Edwardscurve Digital Signature Algorithm (EdDSA) scheme (Ed25519). Cryptographic signing and verification latencies are reported for different message sizes (64 bytes, 256 bytes, and 1024 bytes). Tests demonstrate that Ed25519 achieves the lowest signing latency (36.95-46.53 microseconds) and verification latency (84.55-96.77 microseconds). This equates to processing up to 27,000 signatures per second. Although secp256r1 remains a viable standard, brainpoolP256r1 incurs significant computational overhead, limiting its scalability. Results indicate that Ed25519 is the optimal choice for low-latency V2X authentication, ensuring system reliability even under high vehicle loads.