SegFlow: High-Performance Routing in Payment Channel Networks via Network Partitioning

Abstract

Payment channel networks (PCNs) have emerged as a promising solution for improving the throughput of permissionless blockchains. A key factor affecting PCN performance is routing, which aims to find feasible paths for off-chain payments. However, existing PCN routing schemes either rely on global network knowledge, which causes high overhead and low efficiency, or sacrifice path diversity, which degrades the payment success ratio. In this paper, we conduct a measurement study showing that nodes’ payment behavior exhibits strong locality and further gives rise to stable payment communities in real-world PCNs. Motivated by these observations, we propose SegFlow, a novel routing solution that leverages network partitioning to strike a desirable balance among success ratio, efficiency, and overhead. The core idea of SegFlow is to partition the PCN into multiple subnets and apply differentiated routing strategies to intra-subnet and inter-subnet payments. Specifically, we propose a network partitioning algorithm that enables most payments to be completed within subnets. We then design an inter-subnet routing mechanism that achieves a high success ratio with low overhead for payments across subnets. Extensive simulation and testbed experiments demonstrate that SegFlow achieves a high success ratio comparable to state-of-the-art PCN routing schemes, while substantially reducing overhead and improving routing efficiency.