On Non-Parallelizable Deterministic Client Puzzle Scheme with Batch Verification Modes

A (computational) client puzzle scheme enables a client to prove to a server that a certain amount of computing resources (CPU cycles and/or Memory look-ups) has been dedicated to solve a puzzle. Researchers have identified a number of potential applications, such as constructing timed cryptography, fighting junk emails, and protecting critical infrastructure from DoS attacks. In this paper, we first revisit this concept and formally define two properties, namely deterministic computation and parallel computation resistance. Our analysis show that both properties are crucial for the effectiveness of client puzzle schemes in most application scenarios. We prove that the RSW client puzzle scheme, which is based on the repeated squaring technique, achieves both properties. Secondly, we introduce two batch verification modes for the RSW client puzzle scheme in order to improve the verification efficiency of the server, and investigate three methods for handling errors in batch verifications. Lastly, we show that client puzzle schemes can be integrated with reputation systems to further improve the effectiveness in practice.