On the (Not So) Surprising Impact of Multi-Path Payments on Performance and Privacy in the Lightning Network

The Lightning network (LN) addresses Bitcoin’s scalability issues by providing fast and private payment processing. In order to mitigate failures caused by insufficient channel capacities, LN introduced multi-path payments. To the best of our knowledge, the effect of multi-path payments remains unclear. In this paper, we therefore study the impact of multi-path payments on performance and privacy. We identify metrics quantifying the aforementioned properties and utilise them to evaluate the impact of multi-path payments. To this end, we develop a simulator implementing pathfinding in LN using single and multi-path payments as well as various pathfinding algorithms. We find that, while the success rate of multi-path payments is up to 20% higher, the impact of multi-path payments on performance otherwise remains within limits. On the other hand, the impact on privacy appears to be greater, e.g., multi-path payments are more likely to encounter an on-path adversary and the relationship anonymity is more likely to be compromised by colluding intermediate hops. However, multi-path payments are less likely to be deanonymised based on the path lengths