物联网环境下基于SipHash代理签名的PBFT算法优化

Optimization of PBFT Algorithm Based on SipHash Proxy Signature in the Internet of Things Environment

摘要: 在当前物联网环境中，设备性能与能耗差异显著，加剧了分布式系统一致性问题的复杂性。区块链凭借去中心化共识与可信特性，成为应对该挑战的关键手段。然而，物联网设备常因资源受限、网络波动或电力中断而频繁离线，导致PBFT共识易失败，需频繁重启，从而增加通信开销并降低系统吞吐量。针对上述问题，本文提出一种基于SipHash代理签名的SH-PBFT共识算法，对传统PBFT进行改进。该算法首先设计了一种代理签名节点选择机制，使每个节点均可自主指定代理节点，在自身离线时代为执行签名操作，从而实现代理任务的均衡分配，避免因任务过度集中而导致的性能瓶颈，并构建起一条多级代理签名链，有效提升系统稳定性。其次，本文构建了完整的代理签名模型，涵盖高效安全的签名生成与验证流程，确保签名的有效性、安全性与可追溯性，同时优化了在存在离线节点场景下的共识执行方式，使系统在部分节点离线时仍可完成共识过程，避免因节点失效而触发频繁重启，从而显著降低额外通信时延与开销。最后，仿真实验结果表明，SH-PBFT算法在签名任务分布、系统吞吐量、响应时延及通信开销等方面均表现优良，尤其在节点离线情况下，仍能保持良好的整体性能。

Abstract: In the current Internet of Things environment, there are significant differences in device performance and energy consumption, which exacerbate the complexity of the consistency problem in distributed systems. Blockchain, with its decentralized consensus and trustworthiness features, has become a key solution to this challenge. However, IoT devices often frequently go offline due to resource constraints, network fluctuations, or power outages, leading to the failure of PBFT consensus and frequent restarts, thereby increasing communication overhead and reducing system throughput. To address these issues, this paper proposes an SH-PBFT consensus algorithm based on SipHash proxy signature, which improves the traditional PBFT. The algorithm first designs a proxy signature node selection mechanism, allowing each node to autonomously specify a proxy node to perform the signature operation during its offline period, thereby achieving the balanced distribution of proxy tasks and avoiding performance bottlenecks caused by excessive concentration of tasks. It also builds a multi-level proxy signature chain to effectively enhance system stability. Secondly, this paper constructs a complete proxy signature model, covering efficient and secure signature generation and verification processes, ensuring the validity, security, and traceability of signatures, while optimizing the consensus execution method in the presence of offline nodes, enabling the system to complete the consensus process even when some nodes are offline, avoiding frequent restarts due to node failures, and significantly reducing additional communication delay and overhead. Finally, simulation results show that the SH-PBFT algorithm performs well in terms of signature task distribution, system throughput, response time, and communication overhead, especially in the case of node offline, maintaining good overall performance.