EntrapNet: a Blockchain-Based Verification Protocol for Trustless Computing

Chong Li, Lei Zhang, Serbiao Fang

Abstract

In this paper, we propose a blockchain-based computing verification protocol, called EntrapNet, for distributed shared computing networks, an emerging underlying network for many internet of things (IoT) applications. EntrapNet borrows the idea from the practice of entrapment in criminal law to reduce the possibility of receiving incorrect computing results from trustless service providers who have offered the computing resources. Furthermore, we mathematically optimize EntrapNet to deal with the fundamental tradeoff of a network: security and efficiency. We present an asymptotic optimal solution to this optimization. It will be seen that EntrapNet can be performed as an independent and low-cost layer atop any trustless network that requires outsourced computing, thus making secure computing affordable and practical.

Introduction

As the demand for high-performance computing grows, decentralized networks of idle IoT devices and personal computers offer a scalable alternative to centralized clouds. However, outsourcing computation to "trustless" providers introduces a significant verification bottleneck. Providers may act maliciously or "lazily," skipping computations to save resources. Traditional solutions like task replication, cryptographic proofs (SNARKs), or trusted hardware (TEEs) often suffer from high overhead, hardware dependency, or lack of transparency. EntrapNet addresses these challenges by introducing a lightweight, economic verification layer.

The EntrapNet Protocol

The core mechanism of EntrapNet is inspired by the legal concept of entrapment. The protocol introduces an "Officer" role responsible for maintaining network integrity through the use of "fishing tasks"—trap jobs with pre-verified results.

The EntrapNet architecture separates the system into an Incentive Layer and a Computing Layer, facilitating rewards and deposits to ensure honest behavior.

The workflow follows four key steps:

  1. Task Preparation: The Officer maintains a repository of fishing tasks with known solutions.
  2. Indistinguishable Scheduling: A task scheduler assigns either a regular task or a fishing task to a provider. Crucially, these tasks are indistinguishable to the provider.
  3. Verification and Appeal: If a provider returns an incorrect result for a fishing task, the Officer submits an appeal to a blockchain-based smart contract.
  4. Economic Penalization: Providers must stake a deposit to participate. Upon a failed fishing task, the deposit is forfeited to reward the Officer and replenish the incentive pool.

System Optimization and Performance

A critical challenge in EntrapNet is balancing security (more fishing tasks) with efficiency (fewer redundant tasks). The authors utilize Poisson distribution models to optimize the fishing task rate. To ensure the protocol is computationally feasible for blockchain environments, they developed an "asymptotically optimal" solution.

Numerical simulations show that the mathematical upper and lower bounds for the optimization objective converge toward the true value as the task arrival rate increases.

Simulations based on the Ethereum blockchain demonstrate that this optimized protocol achieves an accuracy of 99.93% compared to the theoretical ideal. Furthermore, the system maintains high security with approximately 10% overhead. The results indicate that as the required security level (deposit size) increases, the optimal number of fishing tasks grows slowly, allowing the network to scale efficiently without sacrificing performance.

Conclusion and Future Work

EntrapNet provides a hardware-independent, low-cost verification layer for trustless computing. By shifting the burden of proof from complex cryptography to economic incentives, it makes decentralized supercomputing more practical. Future research will focus on extending the protocol to non-deterministic tasks (such as machine learning training with floating-point variations), handling massive jobs split across multiple providers, and further hardening the system against sophisticated collusion attacks.

Summary updated

Abstract

In this paper, we propose a blockchain-based computing verification protocol, called EntrapNet, for distributed shared computing networks, an emerging underlying network for many internet of things (IoT) applications. EntrapNet borrows the idea from the practice of entrapment in criminal law to reduce the possibility of receiving incorrect computing results from trustless service providers who have offered the computing resources. Furthermore, we mathematically optimize EntrapNet to deal with the fundamental tradeoff of a network: security and efficiency. We present an asymptotic optimal solution to this optimization. It will be seen that EntrapNet can be performed as an independent and low-cost layer atop any trustless network that requires outsourced computing, thus making secure computing affordable and practical.
BibTeX
@misc{li2021entrapnet,
  author = {Chong Li and Lei Zhang and Serbiao Fang},
  title = {{EntrapNet: a Blockchain-Based Verification Protocol for Trustless Computing}},
  year = {2021},
  eprint = {2105.00146},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2105.00146},
}
Figure