Byzantine-Fault-Tolerant Consensus via Reinforcement Learning for Permissioned Blockchain Implemented in a V2X Network

Abstract

Blockchain has been at the center of various applications in vehicle-to-everything (V2X) networking. Recently, permissioned blockchain gain practical popularity thanks to its improved scalability and diverse needs for different organizations. One representative example of permissioned blockchain is Hyperledger Fabric. Due to its unique execute-order procedure, there is a critical need for a client to select an optimal number of peers. There is a tradeoff in the number of peers: a too large number will lead to a lower scalability and a too small number will leave a narrow margin in the number of peers sufficing the Byzantine fault tolerance (BFT). This channel selection issue gets more due to the mobility: a transaction must be executed and the associated block must be committed before the vehicle leaves a network. To this end, this paper proposes a channel selection mechanism based on reinforcement learning (RL) to keep a Hyperledger Fabric-empowered V2X network impervious to dynamicity due to mobility. We model the RL as a contextual multi-armed bandit (MAB) problem. The results prove the outperformance of the proposed scheme.

On the Byzantine-Fault-Tolerant Consensus for Blockchain among Connected Vehicles

“…2) Related Work: One of this paper's authors has done fundament studies that can be regarded as the basis of this work [2] [3] where the impact of mobility on the performance of block exchange in a vehicular network [2] and the feasibility of permissioned blockchain established on the Hyperledger Fabric [3] were studied.…”

Environment-Adaptive Multiple Access for Distributed V2X Network: A Reinforcement Learning Framework

“…However, feature representation and online learning ability are two major challenges to be solved for learning control of uncertain dynamic systems [13]. As an effort to keep a V2X network's performance stable in such a dynamic environment, a recent work [14] has proposed to apply a MAB-based approach, which turned out to be effective in achieving convergence of learning in a sufficiently short time to deal with the dynamicity. Meanwhile, advanced methods such as federated learning has recently been proposed as a solution to achieve self-adaptation of a wireless system [15]; however, its "localized" validity does not suit our goal of achieving a universal finality.…”