Ripple Protocol performance improvement: Small world theory applied to cross border payments

Abstract

In the context of faster and cheaper international money transfers, this paper makes an analysis of the Ripple Protocol Consensus Algorithm (RPCA) and provides an improved algorithm that takes advantage of the network topology when it holds several sub-networks with poor connections between them. RPCA reaches consensus based on a trusted list of nodes, which this paper theorizes to fit the small world concept from social networks. Network nodes are analyzed statically in order to add new nodes to interconnect clusters. The method used to measure the improvement is through simulation and calculated the algorithm response time with and without the added nodes. This paper achieves a performance improvement between 23 ms and 457 ms for networks of between 50 and 2000 nodes. In the context of high frequency trading, a small milliseconds improvement makes the difference for a transaction to be accepted or not.