Performance Evaluation of zk-SNARK Protocols for Privacy-Preserving Sensor Data Verification: A Systematic...

Area of Science:

• Cryptography
• Computer Science
• Network Security

Background:

• Sensor networks in critical infrastructure, healthcare, and smart cities require privacy-preserving data verification.
• Zero-knowledge succinct non-interactive arguments of knowledge (zk-SNARKs) are a promising cryptographic tool for this purpose.
• The practical deployment feasibility of zk-SNARKs in resource-constrained sensor environments is not well understood.

Purpose of the Study:

• To systematically benchmark the Groth16 zk-SNARK protocol's performance across diverse computational complexities.
• To evaluate the feasibility of using Groth16 for privacy-preserving data verification in sensor networks.
• To provide actionable guidance for designing future sensor network architectures.

Main Methods:

• Benchmarking the Groth16 zk-SNARK protocol using an automated, open-source framework (Circom-snarkjs).
• Testing across eight representative circuit types with computational complexities ranging from 1 to 1,510,185 constraints.
• Conducting 160 statistically controlled measurements, collecting data on proof generation time, verification time, proof size, memory, and witness generation overhead.

Main Results:

• Groth16 proofs exhibit a constant size (804.7±1.7 bytes) and near-constant verification time (0.662±0.032 s) irrespective of circuit complexity.
• Proof generation time scales sub-linearly (α=0.256, R²=0.608) with computational complexity.
• Statistically significant differences in performance were observed across circuit categories (ANOVA: F=355.0, p<10⁻⁷⁹, η²=0.94).

Conclusions:

• The Groth16 zk-SNARK protocol is practically feasible for privacy-preserving data verification in sensor networks due to its consistent proof size and verification time.
• Identified three operational deployment tiers and estimated energy budgets for battery-powered devices.
• Findings offer crucial guidance for designing next-generation, privacy-preserving sensor network systems.