Abstract: This paper presents Dynamic Consensus Optimization (DCO) that integrates online feedback-control framework that augments a Raft-based ordering service with real-time parameter adaptation. DCO continuously monitors transaction arrival rates, queue lengths, and network conditions, and dynamically tunes core Raft parameters such as batching size, election timeouts, and heartbeat intervals without compromising Raft’s safety guarantees. We apply DCO to a simulated healthcare environment, where IoT-enabled devices, sensors, and inter-institution data transfers generate highly variable workloads. The proposed approach yields several advantages. First, the blockchain layer sustains low confirmation latency as IoT workloads vary in volume and complexity across participating institutions. Second, the system scales more gracefully: as the number of member hospitals or patient records grows, DCO dynamically adjusts resources to maintain responsiveness without manual reconfiguration or centralized bottlenecks. Third, decentralization is reinforced through leadership rotation and distributed validation tasks, mitigating centralization of power and reducing single points of failure. By continuously balancing speed, scalability, and safety, DCO enables healthcare providers to adopt blockchain-enabled data sharing with stronger performance guarantees and fewer operational frictions. The paper also outlines a formal argument for safety preservation under dynamic parameter updates and presents an evaluation plan to quantify latency, throughput, scalability, and resilience under heterogeneous IoT workloads.
Siman* et al. (Wed,) studied this question.