Adaptive Routing Mechanism for Transaction Propagation in Wireless Blockchain Network

Traditional routing protocols aim to find the most efficient path between clearly defined source and destination nodes. However, there is no explicit destination node for confirming transactions in the blockchain, due to the inherent uncertainty of the miner who proposes the next block. Existing blockchain systems utilize gossip-based protocols to route transactions, which results in significant communication overhead and impedes the applicability of blockchain to wireless networks. In this paper, we propose an adaptive routing mechanism to achieve efficient transaction routing. This mechanism allows nodes to relay transactions based on forwarding probability rather than direct flooding. Each node uses the no-regret learning algorithm to adjust the probability to avoid invalid forwarding. The proposed routing mechanism significantly reduces bandwidth consumption while ensuring a slight drop in throughput. We prove that our routing mechanism achieves the ε-approximate Coarse Correlation Equilibrium (CCE) and has O(N) time complexity. The results of the simulations demonstrate that the average bandwidth consumption of the proposed routing mechanism is reduced by 90.68% compared to gossip and 45% compared to FISCO BCOS.