Edge Computing and Distributed Intelligence
Résumé
The convergence of Internet of Things (IoT), Artificial Intelligence (AI), 5G, Big Data, and cloud computing technologies is creating an emerging edge to cloud computing model to overcome the challenges of traditional cloud computing in supporting the upcoming Internet of (smart and autonomous) things. Edge computing complements cloud computing by offering local computational facilities to enable processing of extremely large amount of data originated from IoT devices and providing intelligent decisions close to the things. As edge devices are geographically dispersed and may be resource constraint, many challenges exist in deploying and executing intelligent applications over these heterogeneous resources. This chapter will address these challenges through three sections to discuss: (1) what is edge computing and the resource management approaches in edge environments, especially application deployment mechanisms in the IoT-Edge-Cloud continuum; (2) the benefits and challenges of implementing distributed intelligence and data analytics in the edge environments with exemplary use cases; (3) the distributed learning architecture and algorithms for edge computing, with a special focus on distributed deep learning-based approaches.