Abstract
Payment channel networks (PCNs) such as Bitcoin's Lightning Network offer a promising solution for high transaction throughput and near-instant confirmations. However, the network's efficacy depends on participants' willingness to lock collateral into channels. The primary financial incentive for providing this liquidity is the collection of routing fees. Currently, most participants rely on default fee settings, which may not be optimal. This paper investigates the optimization problem for a node operator seeking to maximize financial gain by strategically selecting channel partners and setting fees. We prove that this problem is NP-hard and propose a greedy approximation algorithm. Our simulations, using real-world Lightning Network data, demonstrate that our strategy outperforms standard heuristics by at least a factor of two and increases profits by fourfold compared to default settings.
Introduction
Blockchain scalability remains a significant hurdle for global adoption. Layer 2 solutions, specifically Payment Channel Networks (PCNs), address this by allowing users to conduct transactions off-chain, settling only the final state on the base layer. For a PCN to function, nodes must commit capital to channels, creating an opportunity cost. To offset this, nodes charge fees for routing payments.
While the network's liquidity is crucial, the current incentive structure is underutilized. Most node operators do not strategically choose their connections or fees. This paper addresses the optimization challenge: given a fixed budget, which channels should a node open, and what fees should it charge to maximize its total routing revenue?
Problem Formalization and Complexity
We model the PCN as a weighted directed graph $G = (V, E)$, where edges represent channels and weights represent the fees charged by the nodes. A node $A$ aims to select a set of new edges to create and assign them fees to maximize its total reward, defined by the volume of traffic it routes multiplied by the fees collected.
We demonstrate that this optimization problem is NP-hard through a reduction to the "Maximum Betweenness Improvement" (MBI) problem. This complexity necessitates the use of heuristic or approximation algorithms for practical application in large-scale networks like Bitcoin's Lightning Network.
Methodology
To tackle the computational complexity, we propose a two-part strategy consisting of a Channel Selection Function (CSF) and a Channel Fee Function (CFF).
Channel Selection Function (CSF)
The CSF employs a greedy approach to identify the most profitable channel partners. In each iteration, the algorithm selects the node that provides the highest marginal increase in the expected total reward. This selection is based on "betweenness centrality"—the frequency with which a node appears on the cheapest path between other pairs of nodes in the network. Unlike simple heuristics, this method looks for "gaps" in the network where a node can provide a new, cheaper shortcut.
Channel Fee Function (CFF)
Setting the optimal fee involves a trade-off: higher fees increase the profit per transaction but reduce the likelihood that the node's channel will be part of the cheapest path. We observed a property of "fee monotonicity": as a node increases its fee, its betweenness centrality (and thus the traffic it attracts) can only decrease. Leveraging this, we developed a recursive search algorithm that prunes the search space, finding the optimal fee 10 to 100 times faster than an exhaustive search.
Experimental Evaluation
We evaluated our strategy using a snapshot of the Bitcoin Lightning Network from July 2019, comprising 4,618 nodes and over 68,000 edges. We further refined this data into "medium" and "macro" graphs to account for different transaction sizes and channel capacities.
Key Results
Our simulation results indicate significant improvements over existing practices:
- Profit Maximization: The greedy strategy achieved a 4x increase in profit compared to the default fee settings used by the majority of the network.
- Heuristic Comparison: Our approach outperformed common network-theoretic strategies, such as connecting to nodes with the highest Degree or PageRank, by at least a factor of two.
- Computational Efficiency: The recursive CFF algorithm proved highly efficient, significantly reducing the overhead required to find the "sweet spot" for fees.
Discussion and Limitations
While our results are promising, several factors warrant further investigation. Our current model assumes static channel capacities; in practice, channels can become unbalanced, leading to routing failures. Additionally, we assume a static environment where other nodes do not adjust their fees in response to our strategy. Future work should explore a game-theoretical framework to model the competitive dynamics of rational nodes. Finally, our assumption of a uniform transaction distribution between all node pairs could be refined with more empirical data on actual payment flows.
Conclusion
Strategic participation in PCNs is essential for both individual profitability and the overall health of the network. By formalizing the profit maximization problem and providing an efficient greedy algorithm, we offer a path for node operators to optimize their capital allocation. Our findings suggest that moving away from default configurations can significantly enhance the financial incentives for providing liquidity in decentralized payment networks.
Summary updated