Ant Routing scalability for the Lightning Network
This addresses scaling and decentralization issues in the Lightning Network, an incremental improvement over existing methods.
The paper tackles the scalability of the Ant Routing algorithm for the Lightning Network, proposing an implementation with efficient memory management and proving it can sustain thousands of transactions per second, sufficient for a global payment network.
The ambition of the Lightning Network is to provide a second layer to the Bitcoin network to enable transactions confirmed instantly, securely and anonymously with a world scale capacity using a decentralized protocol. Some of the current propositions and implementations present some difficulties in anonymity, scaling and decentalization. The Ant Routing algorithm for the Lightning Network was proposed in \cite{GrunspanPerez} for maximal decentralization, anonymity and potential scaling. It solves several problems of current implementation, such as channel information update and centralization by beacon nodes. Ant Routing nodes play all the same role and don't require any extra information on the network topology beside for their immediate neighbors. The goal of LN transactions are completed instantaneously and anonymously. We study the scaling of the Ant Routing protocol. We propose a precise implementation, with efficient memory management using AVL trees. We evaluate the efficiency of the algorithm and we estimate the memory usage of nodes by local node workload simulations. We prove that the number of transactions per second that Ant Routing can sustain is of the order of several thousands which is enough for a global payment network.