A Governance Architecture for Self-Adaption & Control in IoT Applications

Abstract

The “Internet of Things” has become a reality with projections of 28 billion connected devices by 2021. Much R&D is currently focused on creating methods to efficiently handle an influx of data. Flow based programming, where data is moved through a network of processes, is a model well suited to IoT. This paper proposes a dynamic, distributed data processing architecture, utilizing a flow based programming inspired approach. We illustrate a distributed configuration management protocol, which coordinates processing between edge devices and a central controller. Our proposed architecture is evaluated in a vehicle use case that predicts driver alertness. We present a scenario for reducing data on vehicular networks when the connectivity options are limited, while maintaining computational accuracy.