Optimization-based Proof of Useful Work: Framework, Modeling, and Security Analysis

TL;DR

This paper proposes a comprehensive framework for optimization-based Proof of Useful Work (PoUW) in blockchain, analyzing its security, efficiency, and designing reward mechanisms to ensure integrity against malicious actors.

Contribution

It introduces a generic PoUW framework using optimization problems, providing security analysis, reward design guidelines, and validation through simulations.

Findings

Security conditions against selfish and malicious miners identified

Trade-off between useful work efficiency and security established

Simulation validates analytical security and performance results

Abstract

Proof of Work (PoW) has extensively served as the foundation of blockchain's security, consistency, and tamper-resistance. However, long has it been criticized for its tremendous and inefficient utilization of computational power and energy. Proof of useful work (PoUW) can effectively address the blockchain's sustainability issue by redirecting the computing power towards useful tasks instead of meaningless hash puzzles. Optimization problems, whose solutions are often hard to find but easy to verify, present a viable class of useful work for PoUW. However, most existing studies rely on either specific problems or particular algorithms, and there lacks comprehensive security analysis for optimization-based PoUW. Therefore, in this work, we build a generic PoUW framework that solves useful optimization problems for blockchain maintenance. Through modeling and analysis, we identify the security conditions against both selfish and malicious miners. Based on these conditions, we establish a lower bound for the security overhead and uncover the trade-off between useful work efficiency and PoW safeguard. We further offer the reward function design guidelines to guarantee miners' integrity. We also show that the optimization-based PoUW is secure in the presence of malicious miners and derive a necessary condition against long-range attacks. Finally, simulation results are presented to validate our analytical results.