Cybersecurity in Large-Scale Smart Cities: Novel Proposals for Anomaly Detection from Edge to Cloud

Abstract

Smart cities are a complicated blend of Information and communications technology (ICT) as well as Internet of Things (IoT) from physical (such as sensors and actuators) and non-physical (such as external databases) data sources. The smart physical data sources contribute to developing the level of safety and security of citizens' daily life through different smart solutions such as optimization for autonomous power grids and human health applications. Those physical devices generate a massive amount of data that is necessary for analytic applications to be applied in user services. Cloud computing technologies utilize as a centralized ICT architecture to collect the massive amount of data from all scale of the city and build the citizens' services by developers. However, several challenges in cybercrime and cybersecurity exist in Cloud computing technologies as well as large scale IoT environments and smart cities. Those challenges are i) the new generation of the attack on an IoT network is to create a problem with the availability of IoT data for the Cloud applications by overwhelming devices with information requests; ii) complex communication and computation resource constraints has a big concern to moving forward to detect attack at the edge of networks. Therefore, heading off the threat of cybercrime in large scale IoT and smart cities calls for a new study on designing cybersecurity models for ICT architecture. In this paper, we first present our fully hierarchal Distributed-to-Centralized ICT (D2C-ICT) architecture for the Zero Emission Neighborhoods (ZEN) center in the context of the smart cities. Second, we propose two different novel anomaly detection models by using the contribution of Fog to cloudlet and Cloud technologies through our presented D2C-ICT Architecture in the context of the smart city. Finally, we discuss two different use cases for attack detection. We envisage that we can enhance the level of attack detection techniques through different technologies of the D2C-ICT from edge to Cloud networks.