A Lightweight DAG-RAFT Hybrid Consensus Framework for Blockchain-Enabled Digital Enterprise Management in Embedded Systems

This manuscript addresses a core challenge in embedded digital enterprise management: reconciling operational efficiency with system security, given the heavy overhead of blockchain on resource-constrained devices. A lightweight Directed Acyclic Graph (DAG)–Reliable, Replicated, Redundant, and Fault-Tolerant (RAFT) hybrid is presented, in which a DAG layer asynchronously orders concurrent transactions through topological relations and predecessor hashes, feeding batched updates to an enhanced RAFT with adaptive heartbeats and dynamic leader election. Version–vector conflict detection and idempotent writes are used to improve log replication and state synchronization. On a 200-node testbed, peak memory and CPU utilization reach 25.9 MB ± 1.8 MB and 20.8% ± 1.6%, respectively; under high load, the system attains 896.5 effective transactions per second (ETPS) with an end-to-end confirmation latency of 203.5 ± 11.2 ms. These results indicate a practical balance of high concurrency, low resource consumption, strong consistency, and robust security for embedded digital enterprise scenarios.