Splitting Payments Locally While Routing Interdimensionally

Lisa Eckey, Sebastian Faust, Kristina Hostáková, Stefanie Roos

Abstract

Payment Channel Networks (PCNs) like the Lightning Network offer a scalable solution for off-chain transactions but are hampered by high failure rates, particularly for large payments. This paper introduces a protocol that empowers intermediaries to split payments locally and dynamically, paired with a novel routing algorithm called Interdimensional SpeedyMurmurs (INTSM). By leveraging multiple spanning trees for routing and homomorphic encryption for transaction unlinkability, the proposed system achieves a success ratio of over 90% for large payments. Crucially, the protocol remains compatible with Bitcoin’s scripting limitations through the use of adaptor signatures.

Introduction: The Reliability Crisis in PCNs

The Lightning Network and similar PCNs rely on source routing, where the sender pre-determines the entire path to the receiver. However, senders lack real-time visibility into the liquid balances of remote channels. If any link in a chosen path has insufficient funds, the entire transaction fails. Simulations indicate that for larger transactions, failure rates can reach between 35% and 54%, a significant barrier to mainstream adoption.

Methodology: Dynamic Local Splitting and INTSM

The authors propose shifting the decision-making power from the sender to the intermediaries, who possess the most accurate information regarding their own local channel capacities.

Local Splitting

Instead of a static path, this protocol allows any intermediary node to split an incoming payment into multiple "sub-payments" on-the-fly. These sub-payments are routed through different neighbors based on real-time liquidity. This dynamic adjustment allows the network to bypass congested or dry channels that a sender could not have predicted.

Interdimensional SpeedyMurmurs (INTSM)

To facilitate efficient routing, the authors developed Interdimensional SpeedyMurmurs. While traditional routing might rely on a single spanning tree, INTSM utilizes multiple trees simultaneously. By calculating distances to the receiver across these multiple "dimensions," nodes gain the flexibility to choose from a wider pool of next hops while still guaranteeing loop-free routing.

Protocol logic for Sender, Intermediary, and Receiver nodes, illustrating the flow of conditional payments and local splitting.

Unlinkability and Security

To prevent intermediaries from discriminating against split payments—which they might do to preserve capacity for simpler transactions—the protocol employs an unlinkability extension. Using additively homomorphic hashing and Paillier-based encryption, the system ensures that intermediaries cannot determine if two incoming sub-payments are part of the same larger transaction. Furthermore, the use of adaptor signatures ensures the protocol is compatible with Bitcoin without requiring a hard fork.

Performance Evaluation

The protocol was evaluated using a March 2020 snapshot of the Lightning Network (6,329 nodes, 84,000+ channels) and various synthetic topologies (Scale-free and Random graphs).

The results demonstrate a significant leap in reliability. For payments averaging 25 EUR, the combination of INTSM and local splitting achieved a success rate exceeding 90%, compared to approximately 77% for standard shortest-path routing. Even without splitting, the INTSM algorithm alone improved success ratios by nearly 50% over baseline protocols. The study also highlighted that "strategic" splitting—prioritizing paths closer to the receiver—is essential, as random splitting can actually degrade performance.

Conclusion and Future Work

By moving splitting decisions to the intermediaries and expanding the routing search space through INTSM, this protocol addresses the fundamental liquidity visibility problem in PCNs. While the current model successfully protects against "Wormhole attacks" and ensures high success rates, future research is required to align routing fee incentives and enhance metadata privacy regarding receiver identities. Long-term studies on how local splitting impacts overall network health and channel depletion will also be vital for sustainable scaling.

Summary updated

BibTeX
@article{eckey2020splitting,
  author = {Lisa Eckey and Sebastian Faust and Kristina Hostáková and Stefanie Roos},
  title = {{Splitting Payments Locally While Routing Interdimensionally}},
  journal = {IACR Cryptol. ePrint Arch.},
  year = {2020},
  url = {https://openreview.net/forum?id=Vw7bcGfIdK},
}
Figure