Smart cities are expected to offer a variety of innovative digital services and to improve quality of life of their citizens. To achieve this, a huge number of sensors permanently monitors different objects generating vast amounts of data which need to be transmitted and processed efficiently. Today’s networks struggle to scale and manage this data. Novel concepts, such a Software Defined Networking (SDN) and Multi-Access Edge Computing (MEC), are introduced to overcome these challenges. In this paper we enhance an IoT architecture for smart cities by integrating SDN for the optimal allocation of smart city applications between MEC centers and the cloud. This decision is taken using machine learning based on historical and current data. Some smart city applications illustrate this approach.
Paper
Full text
IoT Architecture for Smart Cities Leveraging Machine Learning and SDN
Semantic Scholar · Computer Science · 2019
Abstract
Smart cities are expected to offer a variety of innovative digital services and to improve quality of life of their citizens. To achieve this, a huge number of sensors permanently monitors different objects generating vast amounts of data which need to be transmitted and processed efficiently. Today’s networks struggle to scale and manage this data. Novel concepts, such a Software Defined Networking (SDN) and Multi-Access Edge Computing (MEC), are introduced to overcome these challenges. In this paper we enhance an IoT architecture for smart cities by integrating SDN for the optimal allocation of smart city applications between MEC centers and the cloud. This decision is taken using machine learning based on historical and current data. Some smart city applications illustrate this approach.