The realization of wireless communication in rural environments suffers from long distances, unknown node mobility and discontinuous communication paths. Combination of delay-tolerant mobile ad-hoc networks and infrastructure-based mobile communications results in an increased number of communication opportunities in many usage scenarios. However, efficient exploitation of communication opportunities remains a challenging task.

This article introduces a novel middleware approach specialized for consistent and efficient wireless communication in rural environments. An important part of communication optimization is realized by processing on-hand information of the current application scenario. The middleware autonomously links scenario information to network states and applies machine-learning-based analyses to derive meaningful communication predictions. Cooperation of infrastructure-based and ad-hoc communication is implemented by a gateway component. The gateway collects connectivity information of mobile nodes, enhances communication predictions and distributes predictions across the network. Evaluation of local communication layer and application layer information as well as sharing of communication predictions using an application level API ensure efficient use of communication opportunities.