Adaptive Fog Computing Architecture for Scalable Smart City Infrastructure

Abstract

The hypergrowth in urban areas and the exponential growth in the number of the Internet of Things (IoT) devices have resulted in the unprecedented growth of the production of the data in real-time in many areas including traffic control, environmental monitoring, public safety, and energy management. Conventional cloud-centric computer systems have the benefit of significant scale in such aspects as latency, bandwidth bottlenecks and contextual awareness; however, such systems are inappropriate to satisfy latency-sensitive applications such as smart city environments. In order to tackle such a challenge, this paper proposes a new Adaptive Fog Computing Architecture (AFCA) which can be used to provide a scalable, low-latency termed Adaptive Fog Computing Architecture (AFCA) context-aware system of managing the heterogeneous and geographically spread data sources in cities. AFCA adds a three tier structure of computing that combines the edge, fog, and cloud tiers, which are complemented with two major advancements: a Context-Aware Decision Engine (CADE) and a Resource Orchestration Layer (ROL). CADE and ROL minimize and backward propagates the delays with a dynamic priority assignment and routing algorithms with consideration to latency-sensitive, load, and service criticality. ROL adopts AI-based forecasting model to distribute the computational resources dynamically among fog nodes. Provided system architecture will be tested in a hybrid testbed with: Raspberry Pi edge nodes, Intel NUC fog servers and AWS-hosted cloud instances. Realistic smart city benchmarks involving smart parking, pollution sensing and traffic monitoring datasets indicate that AFCA can reduce processing latency by 42%, experience 31% higher energy efficiency and 45% more device handling capacity than conventional fogs. Also, the architecture follows the use of lightweight security mechanisms and differential privacy approaches to prevent data integrity and user privacy. It is highly modular and predictive, which makes it conveniently work in fast-paced urban environments where the demand of the services and the condition of the network constantly change. On the whole, AFCA constitutes an interesting innovation in fog computing paradigms, and is an ideal, flexible platform on which to realize the deployment of smart services within the infrastructures of the next generation of smart cities.