Smart cities are defined as urban areas that integrate digital technologies to improve the quality of services and optimize resource management. IoT, big data analytics, Artificial Intelligence (AI), and cloud computing form the backbone of smart city infrastructure. These technologies allow for data-driven decisions, real-time monitoring, and automation of city functions. For example, IoT sensors in smart cities are employed to track air quality, manage energy consumption, and monitor traffic patterns in real-time [3].

Cloud computing is central to processing and storing the massive amounts of data IoT devices generate. However [4], cloud-centric models often fail to meet the low-latency requirements of many smart city applications, such as emergency services, smart grids, or autonomous vehicle systems. Furthermore, transmitting all data to the cloud increases network congestion and energy consumption. Edge computing addresses these issues by processing data closer to the source, reducing the load on centralized cloud systems, and enhancing real-time decision-making.

In a smart city, data from thousands or millions of IoT devices, such as sensors, cameras, and GPS units, must be collected, processed, and analyzed. Cloud computing offers a centralized platform for handling this data, providing storage, computational power, and data analytics tools through services like Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) [5].

The cloud enables cities to scale their data processing infrastructure according to demand. For example, more computational resources can be allocated during peak traffic hours to process traffic data, optimize signal timings, and minimize congestion. Cloud platforms also allow cities to implement advanced AI and machine learning algorithms to analyze historical data to predict future trends, such as energy demand or transportation needs.

Despite its benefits, cloud-centric models are not without their limitations [6]. These models require all data from IoT devices to be transmitted to a centralized cloud server for processing. This transmission creates several challenges:

I. High latency: In applications where real-time data processing is critical, such as traffic management or public safety systems, delays caused by sending data to and from the cloud can result in inefficiencies or even failures in service delivery.

II. Network congestion: As the number of IoT devices grows, the amount of data transmitted to the cloud increases, resulting in network congestion. This congestion can further exacerbate latency and reduce the quality of service.

III. Energy consumption: Continuously transmitting data over long distances consumes significant energy, both in IoT devices and network infrastructure. This energy consumption increases operational costs and makes cloud-centric models less sustainable for large-scale smart city deployments.

Cloud computing has enabled large-scale smart city initiatives in cities like Singapore and New York. For example, Singapore's Smart Nation program uses cloud computing to integrate data from traffic systems, energy grids, and public services to provide a seamless urban experience [7]. However, even in these advanced systems, the limitations of cloud-based architectures have led to the exploration of edge computing solutions to enhance real-time processing capabilities and reduce energy costs.

Traditional IoT frameworks rely heavily on centralized cloud computing, which creates several bottlenecks as the number of connected devices grows:

Scalability: Managing millions of IoT devices within a city requires a highly scalable infrastructure. Traditional cloud-centric frameworks struggle to scale effectively due to the centralized nature of data processing. As a result, systems become overloaded, leading to increased latency and decreased performance.

Energy efficiency: IoT devices face significant energy constraints, particularly those with limited battery life. Transmitting data to the cloud for processing requires considerable energy, reducing the devices' operational lifespan.

Security and privacy: IoT devices continuously collect and transmit sensitive data, such as location, personal information, and usage patterns. Ensuring the security and privacy of this data is critical to maintaining public trust in smart city initiatives.

This paper proposes a hybrid cloud-edge computing framework for smart city IoT deployments to address these challenges. In this architecture, edge devices, such as gateways or local servers, handle time-sensitive data processing tasks closer to the source. Meanwhile, the cloud is used for tasks that require greater computational power, such as long-term data storage and advanced analytics.

The hybrid framework offers several advantages:

Reduced latency: By processing data locally at the edge, the system can reduce the time it takes to make critical decisions, such as adjusting traffic signals in response to real-time traffic conditions.

Improved scalability: Edge computing's distributed nature allows the system to scale more efficiently, as each edge node can handle a subset of the overall data load. This reduces the strain on the central cloud server and improves system performance.

Energy efficiency: Edge devices consume less energy by processing data locally, reducing the need for continuous data transmission to the cloud. Additionally, energy-efficient algorithms can be implemented at both the cloud and edge levels to reduce power consumption further.

Current cloud architectures for smart cities are often centralized, with all data from IoT devices being transmitted to a cloud server for processing. This approach works well for applications that do not require real-time data processing, such as long-term data analysis or storage. However, many smart city applications require near-instantaneous responses, which centralized architectures cannot provide due to the following limitations:

Latency: The time taken to transmit data from IoT devices to a central cloud server and back introduces unacceptable delays for time-sensitive applications, such as emergency response systems or autonomous vehicles.

Bandwidth limitations: Transmitting all IoT data to the cloud consumes significant bandwidth, which can result in network congestion, especially in cities with large populations and high IoT device density. This congestion can further exacerbate latency issues.

Energy consumption: Continuously transmitting data from IoT devices to the cloud consumes significant energy, both in the devices themselves and the network infrastructure. This energy consumption is particularly problematic for battery-powered devices, such as environmental sensors with limited operational lifespans.

To address these challenges, this paper proposes a cloud-edge hybrid architecture that combines the strengths of both cloud and edge computing. In this architecture, edge devices are responsible for processing time-sensitive data locally, while the cloud handles more complex tasks, such as data storage and advanced analytics. The hybrid architecture has several key components:

Edge devices: These devices are located close to the IoT sensors and handle data processing tasks that require low latency, such as real-time traffic management or public safety systems. By processing data locally, edge devices reduce the need for continuous data transmission to the cloud, thereby reducing network congestion and energy consumption.

Cloud server: The cloud server handles tasks that require significant computational power, such as long-term data storage, machine learning model training, or large-scale data analytics. By offloading these tasks to the cloud, the system can take advantage of the cloud's scalability and computational resources.

Data aggregation and distribution: The hybrid architecture includes a layer of data aggregation and distribution, which ensures that data is efficiently routed between the cloud and edge devices. This layer also includes security protocols to ensure that data is transmitted securely between the various system components. The proposed architecture is designed to be flexible and scalable, allowing cities to adapt their IoT infrastructure to meet the specific needs of different applications. For example, edge devices can handle real-time energy consumption monitoring in a smart grid system, while the cloud server can perform long-term analysis to predict future energy demand.

The energy consumption of IoT systems is a significant concern, particularly in applications where devices must operate for extended periods without access to a power source. For example, environmental sensors used to monitor air quality in a smart city may be deployed in remote locations where providing a continuous power supply is difficult or expensive. Similarly, smart parking sensors that monitor the availability of parking spaces must operate for long periods on limited battery power.

In traditional cloud-centric models, IoT devices must continuously transmit data to a central server for processing. This continuous transmission consumes significant energy, particularly in wireless networks where data transmission is energy-intensive. As a result, the battery life of IoT devices is often significantly reduced, leading to increased maintenance costs and decreased system reliability.

To address these challenges, this paper proposes using energy-efficient algorithms for cloud and edge computing in smart cities. One such algorithm is Self-adaptive Particle Swarm Optimization (SPSO), which dynamically allocates computational tasks between the cloud and edge devices based on real-time energy constraints [10].

SPSO works by continuously monitoring IoT devices' energy consumption and adjusting task distribution accordingly. For example, if an edge device is running low on battery power, the system can offload more tasks to the cloud, extending the device operational lifespandevice's operational lifespan. Conversely, when the edge device has sufficient power, it can handle more data processing tasks locally, reducing the need for data transmission and conserving energy.

The algorithm also incorporates predictive analytics to forecast future energy demand and adjust the system's operations accordingly. For example, in a smart grid system, the algorithm can predict periods of high energy demand and adjust the distribution of computational tasks to ensure that critical systems remain operational during peak usage periods.

In addition to energy-efficient algorithms, this paper explores the potential of energy-harvesting technologies to power IoT devices in smart cities. Energy harvesting refers to the process of capturing and storing energy from the surrounding environment, such as solar power, wind energy, or traffic vibrations.

Energy harvesting technologies can significantly extend the operational lifespan of IoT devices by providing a continuous power source. For example, solar-powered environmental sensors can operate for extended periods without battery replacement, reducing maintenance costs and improving system reliability. Similarly, piezoelectric sensors can harvest energy from vibrations in roads or bridges to power smart transportation systems.

IoT systems are inherently vulnerable to various security threats due to their distributed nature and many connected devices. Some of the key security challenges include:

Data breaches: IoT devices continuously collect sensitive data, such as personal information, energy consumption patterns, and transportation data. If unauthorized parties intercept this data, it can lead to significant privacy violations.

Denial of Service (DoS) attacks: IoT devices are vulnerable to DoS attacks, in which malicious actors overload the system with traffic, making it unresponsive. These attacks can disrupt critical smart city services, such as traffic management or emergency response systems.

Device tampering: IoT devices are often deployed in public spaces, making them vulnerable to physical tampering. For example, an attacker could tamper with a smart parking sensor to provide false information about the availability of parking spaces.

This paper proposes a multi-layered security framework for IoT systems in smart cities to address these challenges. The framework incorporates several layers of security, including encryption, secure communication protocols, and AI-based threat detection systems.

Encryption: All data transmitted between IoT devices and the cloud is encrypted to prevent its interception by unauthorized parties. This includes data at rest (stored data) and in transit (data transmitted across the network).

Secure communication protocols: The framework employs secure communication protocols, such as Transport Layer Security (TLS), to ensure that data is transmitted securely between IoT devices, edge devices, and cloud servers.

AI-based threat detection: The framework incorporates AI-based threat detection systems that continuously monitor network traffic for signs of malicious activity. These systems can detect and respond to threats in real-time, ensuring that the system remains secure despite sophisticated attacks.

In addition to security measures, the framework incorporates privacy-preserving techniques to ensure that personal data remains protected throughout its lifecycle. One such technique is data anonymization, which removes personally identifiable information from the data before it is transmitted to the cloud. This ensures that even if the data is intercepted, it cannot be used to identify individuals.

The framework also includes privacy-preserving data aggregation techniques, which allow the system to analyze data without revealing individual-level information. For example, a smart energy system can analyze aggregate energy usage patterns without accessing the energy usage data of individual households.

Barcelona has implemented an advanced traffic management system that uses IoT sensors and cloud computing to monitor and optimize traffic flow in real time. The system uses sensors embedded in roads, traffic lights, and vehicles to collect data on traffic patterns, vehicle speeds, and congestion levels.

This data is transmitted to a cloud server, where it is analyzed using machine learning algorithms to predict traffic congestion and optimize traffic signal timings. The system can also detect accidents or traffic jams in real time and adjust traffic signals to reroute vehicles and minimize delays.

In addition to traffic management, Barcelona has implemented an IoT-enabled energy management system to optimize renewable energy sources. The system uses IoT sensors to monitor energy consumption in real-time and adjust energy distribution based on demand.

For example, during periods of high energy demand, the system can prioritize using renewable energy sources, such as solar or wind power, to reduce the city's reliance on fossil fuels. The system can also predict future energy demand based on historical data and adjust the energy distribution accordingly.

Barcelona's case study demonstrates the effectiveness of cloud-IoT frameworks in managing urban infrastructure. However, it also highlights the limitations of centralized cloud computing in handling real-time data processing. By integrating edge computing into its IoT framework, Barcelona has reduced latency and improved the responsiveness of its traffic management and energy systems.