Figure 1:
Figure 2:
Figure 3:
Figure 4:
Figure 5:
Figure 6:
Comparative table
| Reference | Key findings | Advantages | Disadvantages | Accuracy (%) | Remarks |
|---|---|---|---|---|---|
| Smith et al. (2020) | AI-driven firewalls system improves anomaly detection in smart networks | Real-time threat detection | Excessive cost of implementation | 92 | Effective but expensive solution for the large cities’ landscape |
| Chen and Huang (2021) | AI firewall system reduces energy wastage in IoT networks by optimizing data flows | Energy-efficient | Scalability concerns | 89 | Suitable for small-scale real-time applications |
| Gupta and Kumar (2021) | Highlighted the data privacy risks involved with AI firewalls | Advanced threat prevention | Potential data privacy issues | 88 | Essential for data-sensitive cities |
| Lee et al. (2021) | Compared the intelligence firewalls with traditional classical firewalls in smart transportation systems | Improved adaptability | Complexity in implementation | 94 | Highly adaptable for transport networks |
| Al-Sharif et al. (2021) | AI firewalls enhance response times for cyber threats in public reliance networks | Faster threat response | Requires skilled maintenance personnel | 90 | Effective critical response systems |
| Patel and Sinha (2022) | AI-based firewall system protects against DDoS attacks in smart health care systems | Reduces downtime | High initial investment | 93 | Vital for health care infrastructure |
| Wang et al. (2022) | Examined energy consumption of AI-driven firewalls in environmental monitoring | Energy-efficient | Limited to specific data types | 87 | Ideal for IoT environmental applications |
| Singh and Zhang (2022) | AI firewall system improves security and data accuracy in autonomous vehicles | High detection accuracy | Costly hardware requirements | 95 | Promising autonomous systems in smart cities |
| Rahman and Ali (2022) | AI-driven predictive analytics for identifying vulnerable points in smart city landscapes | Proactive threat mitigation | Limited scalability | 90 | Great for proactive city cyber defense |
| Martin and Lee (2023) | AI firewalls enhance the scalability of security solutions in urban infrastructures | Scalable across networks | High computational needs | 92 | Useful for large urban areas |
| Gomez et al. (2023) | Evaluated the impact of AI firewall systems on protecting public IoT devices in parks and public spaces | Enhanced protection | Frequent updates needed | 88 | Suitable for public IoT devices |
| Lee and Nakamura (2023) | Analyzed the adaptability of AI-driven firewalls in urban traffic systems | Real-time adaptability | Expensive maintenance | 91 | Effective for adaptive traffic control |
| Hussein and Omar (2023) | AI firewalls contribute to low energy consumption in smart grids, enhancing sustainability | Energy-efficient | May affect data latency | 89 | Best for smart energy networks |
| Malik and Rahim (2023) | Highlighted data privacy concerns when using AI-driven firewalls for citizen information | Secures personal data | Data privacy challenges | 87 | Vital for privacy-focused cities |
| Brown et al. (2023) | Examined the effectiveness of the AI firewall system in secure public safety communications | Enhanced data protection | High technical expertise required | 94 | Essential for emergency response |
| Hassan et al. (2024) | Presented AI firewalls reduce cyberattack risks in smart city waste management systems | Reduces operational disruption | Potential excessive cost | 90 | Ideal for waste management |
| Lee and Chen (2024) | Evaluated integration challenges of the AI firewall system across heterogeneous IoT devices in smart cities | High compatibility | Integration complexities | 86 | Requires standardization for broader use |
| Xu and Park (2024) | Compared AI firewall response time to traditional systems in managing cyberattacks on energy systems | Faster response time | Limited to specific devices | 93 | Promising energy infrastructure |
| Jung and Lee (2024) | Discussed self-learning AI-driven firewalls that autonomously adapt to new cyber threats in real-time | High adaptability | Requires continuous data updates | 96 | Effective in constantly evolving networks |
| Chen et al. (2024) | Reviewed AI-driven firewall applications in managing secure data flow in transportation systems, highlighting its impact on reducing energy and data processing costs | Cost-effective | High installation cost | 92 | Ideal for energy-efficient transport systems |
Comparative study of related works and current research on AI firewalls for smart cities
| Aspect | Previous work | Current work (smart cities secured) | Research gaps addressed |
|---|---|---|---|
| Focus | Emphasized the vulnerabilities of IoT devices and traditional classical firewalls in smart cities (Zhu et al., 2021; Patel & Sinha, 2022) | It explores AI-driven firewall systems as adaptive and robust solutions for cybersecurity in smart cities | By highlighting the transformative role of AI firewall systems in combating evolving cyber threats |
| Technological approach | Investigated traditional classical firewalls and initial machine-learning models for anomaly detection (Chen et al., 2024) | Introduces advanced AI algorithms, such as self-learning and predictive analytics, for real-time threat management | Demonstrates how advanced AI capabilities are superior to traditional methods in smart city applications |
| Cybersecurity challenges | Prioritizing individual vulnerabilities in IoT systems without a comprehensive framework (Yin et al., 2020) | Provides a comprehensive approach to addressing interconnected vulnerabilities across different smart city services | Integrates fragmented insights into an integrated AI powered cybersecurity approach |
| Sustainability impact | Limited focus on the cybersecurity and sustainability relationship (Hassan et al., 2024) | Connects robust cybersecurity measures with sustainable urban development objectives, utilizing efficient energy utilization | Expands the discussion to include the distinct advantages of security and environmental sustainability |
| Implementation challenges | Highlighted excessive cost and data privacy concerns as obstacles to adoption (Malik & Rahim, 2023) | Discusses solutions such as federated learning and edge AI-driven firewalls to overcome cost, scalability, and privacy concerns | Provides innovative approaches to tackle financial and technical issues |
| Future trends | Scalability and interoperability were identified as emerging cyber issues but provided limited solutions (Lee & Chen, 2024) | It explores emerging trends such as XAI and edge computing for improved scalability and transparency | Creates effective strategies for future research and real-world implementation |
AI firewall metrics for smart city security and efficiency
| Metric | Example data | Description | Data source |
|---|---|---|---|
| Detection accuracy | 97.5%–98% | AI firewall systems have shown high detection precision in identifying cyber threats in real-time systems, including DDoS attacks and unusual traffic patterns | London AI-driven security system pilot project |
| Response time | <1 s (0.8 s) | The response time of AI firewalls to eliminate and eliminate threats is extremely rapid, ensuring minimal impact from cyberattacks | New York City smart city utility environment |
| Energy efficiency | 30% reduction in energy consumption | AI-driven firewalls optimize energy utilization by adjusting resources during low-risk periods, eliminating the energy consumption of monitoring systems without compromising security | New York City data Centre energy optimization case study |
| Proactive threat detection | 35% improvement in cybersecurity posture, detecting cyber threats two weeks in advance | AI-driven firewalls utilize ML models to predict vulnerabilities and detect cyber threats proactively, allowing cities to mitigate potential risks | Singapore predictive anomaly detection system |
| Cost efficiency | 25% reduction in cybersecurity-related costs over the first year | AI firewalls reduce cybersecurity costs, including service downtime and manual intervention, leading to greater cost-effectiveness | Barcelona Smart City Initiative Scheme |
| Scalability and adaptability | 500,000+ devices covered without performance degradation | The AI firewall systems’ ability to scale encompasses many devices across various urban areas, maintaining consistent performance and reliability | Singapore smart city network scalability tests |
Traditional firewalls vs AI firewalls
| Feature | Traditional firewalls | AI firewalls |
|---|---|---|
| Threat detection | Relies on predefined rules and signatures | Uses ML and DL algorithms to detect new and unknown threats based on data patterns |
| Adaptability | Static rules require manual updates for new threats | Continuously learn and adapt to evolving threats |
| Real-time response | Limited real-time capabilities | Can autonomously detect and mitigate threats in real-time |
| Scalability | May struggle with large, dynamic IoT environments | Scalable; adapts seamlessly as new IoT devices are integrated into the network |
| Response speed | Slower response time; human intervention is often required | Instant response with minimal human intervention |
| Anomaly detection | Limited to known attack signatures | Detects new anomalies and behavioral changes in real-time |
| Resource efficiency | May cause delay due to rule-based operations | Optimized for resource efficiency, using ML to process data efficiently |
| Deployment complexity | It is easier to deploy but requires regular manual updates | It is more complex to deploy but provides automated and long-term security benefits |