Multi-rail routing of incoming network data

Abstract

A method for learned multi-rail routing of incoming data includes: generating, by a model training application executing on a processor, a data multi-rail routing model by: receiving, collecting, and accessing historical data corresponding to at least one transfer event, where each transfer event is associated with a source, a source endpoint, and a destination endpoint; training the routing model using the historical data to build a processing plan for each incoming transfer event including identifying a processing route selected from a plurality of rails; and testing the trained routing model; deploying the trained routing model to a processing application for processing incoming transfer events; receiving, by a processing application executing on a processor, transfer event initiation data; and determining, by the deployed model executing on a processor based on the initiation data, a type of transfer event to associate with the initiation data.