Fog Computing Approaches in Smart Cities: A State-of-the-Art Review (2011.14732v2)

Abstract: These days, the development of smart cities, specifically in location-aware, latency-sensitive, and security-crucial applications (such as emergency fire events, patient health monitoring, or real-time manufacturing) heavily depends on a more advance computing paradigms that can address these requirements. In this regard, fog computing, a robust cloud computing complement, plays a preponderant role by virtue of locating closer to the end-devices. Nonetheless, utilized approaches in smart cities are frequently cloud-based, which causes not only the security and time-sensitive services to suffer but also its flexibility and reliability to be restricted. So as to obviate the limitations of cloud and other related computing paradigms such as edge computing, this paper proposes a systematic literature review (SLR) for the state-of-the-art fog-based approaches in smart cities. Furthermore, according to the content of the reviewed researches, a taxonomy is proposed, falls into three classes, including service-based, resource-based, and application-based. This SLR also investigates the evaluation factors, used tools, evaluation methods, merits, and demerits of each class. Types of proposed algorithms in each class are mentioned as well. Above all else, by taking various perspectives into account, comprehensive and distinctive open issues and challenges are provided via classifying future trends and issues into practical sub-classes.

• The paper establishes a taxonomy of fog computing approaches using a systematic literature review of 32 key studies from 2013 to 2020.
• It demonstrates how distributed computing near end-devices reduces latency and enhances real-time applications like emergency response and traffic management.
• It identifies research gaps in security, resource optimization, and real-world validations, urging development of more efficient (meta-)heuristic algorithms.

This paper, "Fog Computing Approaches in Smart Cities: A State-of-the-Art Review" (2011.14732), presents a Systematic Literature Review (SLR) analyzing how fog computing is being applied in smart city contexts, focusing on research published between 2013 and 2020. It addresses the limitations of cloud-only architectures for smart cities, particularly for applications requiring low latency, location awareness, and enhanced security, such as emergency response, real-time traffic management, and patient monitoring. Fog computing, by distributing computation, storage, and networking functions closer to end-devices (IoT sensors, vehicles, etc.), is positioned as a necessary complement to the cloud.

Methodology:

The review follows a structured SLR process, starting with 726 papers identified through a targeted search string ("fog" AND ("city" OR "cities" OR "urban")) across major academic databases. Applying specific inclusion/exclusion criteria (focusing on peer-reviewed papers detailing techniques, solutions, or assessments specifically for fog in smart cities), the authors selected 32 primary studies for detailed analysis. The methodology aimed to be more comprehensive than previous reviews by explicitly including the term "urban" and differentiating fog from edge computing.

Key Findings and Classification:

The core contribution is a taxonomy classifying the 32 reviewed papers into three main categories based on their primary focus:

1. Service-Based Approaches (approx. 43% of papers): These studies focus on developing or improving specific services within a smart city using fog computing.
   • Sub-classes: Data management (e.g., hierarchical architectures for big data analysis), Traffic management (e.g., vehicular fog computing for routing and congestion mitigation), Energy management (e.g., using meta-heuristics to balance load and energy), Monitoring (e.g., distributed video surveillance summarization), and Security (e.g., middleware for integrating fog and Cloud of Things).
   • Example Implementations: Architectures like F2c2C-DM for hierarchical data management, FOREVER for traffic management, and SmartCityWare middleware.
2. Resource-Based Approaches (approx. 28% of papers): These studies concentrate on the management of fog computing resources themselves.
   • Sub-classes: Scheduling (e.g., adaptive genetic algorithms for task scheduling), Load balancing (e.g., distributed algorithms based on the Hungarian method), Offloading (e.g., secure computation offloading using machine learning like Neuro-Fuzzy and PSO), Resource Allocation (e.g., multi-layer architectures like FOCAN), and Resource Provisioning (e.g., fog-cloud coordination strategies).
   • Example Implementations: Algorithms like ADGTS for task scheduling, SecOFF-FCIoT for secure offloading, and architectures like FOCAN.
3. Application-Based Approaches (approx. 28% of papers): These studies explore specific end-user applications enabled or enhanced by fog computing in smart cities.
   • Sub-classes: Vehicle/IoV (e.g., mobility support frameworks, sensing platforms like SensingBus), Smart lighting (e.g., SSL architecture for energy-efficient street lamps), Smart buildings (e.g., integrated IoT-fog-cloud architectures), Noise detection systems (e.g., scalable frameworks for urban sound classification), Hospitality industry, and Self-regulating systems.
   • Example Implementations: Frameworks for IoV, the SensingBus prototype using city buses as mobile fog nodes, SSL architecture for smart streetlights.

Analysis and Insights:

• Evaluation Metrics: Latency (evaluated in 26% of studies) and energy consumption (18%) are the most frequently assessed metrics, reflecting key motivations for using fog. However, crucial aspects like security (4%) and throughput (5%) appear significantly under-evaluated. Scalability (16%) and response time (12%) receive moderate attention.
• Evaluation Methods: Simulation (59%) is the overwhelmingly dominant evaluation method. Real-testbed implementations (16%) and prototypes (19%) are less common, indicating a gap between theoretical proposals and validated real-world deployments.
• Algorithms: Non-heuristic algorithms (44%) are most prevalent, followed by heuristic (31%) and meta-heuristic (25%) algorithms, often applied to optimization problems like scheduling and load balancing.
• Tools: Common tools and platforms mentioned include FIWARE, OpenStack, Java, Hadoop, Python, NS-2/NS-3, SUMO, MATLAB, and databases like MySQL and MongoDB.

Open Issues and Future Directions:

The review identifies several critical areas for future research and practical implementation:

1. Optimal Algorithms: Developing more efficient (meta-)heuristic algorithms for NP-hard problems common in fog environments (scheduling, routing, resource allocation).
2. Big Data Analytics: Better leveraging fog's proximity for real-time processing and analysis of the massive, diverse data generated by smart cities.
3. Green Computing: Focusing on energy optimization strategies for sustainable smart cities.
4. Social Network Integration: Utilizing data from Online Social Networks (OSNs) for resource prediction, service demand analysis, and management in smart cities.
5. Application Deep Dives: Further research within specific domains like autonomous transportation (IoV), smart homes, smart healthcare, smart surveillance, and smart parking/shopping, addressing their unique challenges.
6. Security and Privacy: This remains a major, under-addressed challenge requiring robust solutions for data protection, access control, and preventing misuse across distributed fog nodes.
7. Mobility Management: Developing effective strategies to handle mobile IoT devices and fog nodes (e.g., in vehicles), enabling seamless service continuity and context-awareness.
8. QoS and Scalability: Designing multi-objective optimization techniques for QoS trade-offs and rigorously evaluating solutions for large-scale, real-world deployment scalability.
9. Real-World Validation: Moving beyond simulation to more real-testbed deployments and prototype validation to understand practical challenges and adaptability.

In conclusion, the paper provides a structured overview of the state-of-the-art in fog computing for smart cities, highlighting its potential benefits, classifying existing approaches, analyzing current research trends (especially regarding evaluation), and outlining significant open challenges and practical considerations for future implementations.