Software-defined internet of things for smart urban sensing

Abstract

With more people living in cities, urban sensing is urgently required to create a comfortable and convenient living environment. As Internet of Things (IoT) is the fundamental infrastructure to realize urban sensing, it should be flexible to support various application requirements and convenient management of infrastructure. Inspired by software-defined networking, which aims to make networks more flexible, the authors propose a software-defined IoT architecture for smart urban sensing. This architecture decouples urban sensing applications from the physical infrastructure. Centralized controllers are designed to manage physical devices and provide APIs of data acquisition, transmission, and processing services to develop urban sensing applications. With these properties, various applications can coexist on the shared infrastructure, and each application can request controllers to customize its data acquisition, transmission, and processing on-demand by generating specific configurations of physical devices. This article discusses the background, benefits, and design details of the proposed architecture as well as open problems and potential solutions to realize it, which opens a new research direction for IoT and urban sensing.