Article Preview
TopIntroduction
The number of the growing Internet of Things (IoT) based smart devices produces a massive volume of data that need immediate processing of information. IoT users will store and process their information domestically on their near processing nodes, i.e., fog nodes. Fog nodes (i.e., small scale datacenters) are deployed natively to end-users to store and process their information domestically faster than the cloud (Jalali, 2016) Fog computing facilitates the users with low latency, localization, high-rate services (Deng, 2016).
Figure 1 shows the fog computing network architecture, which contains three layers, i.e., IoT layer, which have intelligent devices that communicate with each other and fog nodes in Fog Node (FN) layer. Fog nodes are further connected to the fog small scale data servers.
Figure 1. Fog computing network architecture
More the number of IoT users, there will be more the resource requirement, and there is a need to utilized these resources for which load balancing is required. Sometimes only a few resources are being used for processing user requests; they keep on receiving and processing information. Other resources in the system remain under-utilized. Load balancing is required to avoid under-utilization of resources, which will help distribute the workload among all the resources equally. More the resource requirement more will be energy consumption in fog nodes. A few latency-sensitive applications like intelligent healthcare, intelligent traffic management, smart city, and fog computing provide alternate solutions for cloud services (Saroa, 2018) (Wadhwa, 2018). However, besides these services, fog computing small scale datacenters consumes a large quantity of energy and emits carbon and different gasses harmful to the surroundings. The fog layer's workload can be defined as the total number of tasks assigned to the fog nodes amongst the available jobs. In innovative applications like traffic management, smart city, there is a tremendous amount of data generated needing more processing resources. In this scenario, energy consumption is more because more resources consume much power. Therefore, to safeguard the surroundings, there is a requirement to develop simple solutions that may scale back the price and reduce harmful effects on the environment.
TopIssues Faced By Traditional Traffic Management Systems
Traditional systems such as vehicular Adhoc network has to face some issues like latency, more energy consumption, high implementation cost. Fog computing has a broad scope in intelligent applications. Fog computing tries to resolve such problems of traditional systems. In urban areas, vehicular networks are considered as the main components of intelligent transportation systems, which covers different areas like road safety, navigation and localization of vehicles, and information spreading. According to one study, it has been estimated that since 2020 there have been 150 million vehicles that are connected on roads (Ning Z. a., 2019) . According to Huang, C. et al. (Huang, 2017), one car generates 30 terabytes of data on average per day. With the increase in the number of vehicles per day, the data generation rate in intelligent traffic management systems also increased. In modern smart traffic management systems, with increased data generation, more resources are required to handle these requests. So, computing resources may suffer from overloading requests that need immediate processing, i.e., in case of medical emergencies. Traditional traffic systems also face low quality of services and low latency while communicating between different vehicles (Wu, 2020). So, there is a need to generate an intelligent traffic management system that can tackle problems faced by traditional systems. Although there are some solutions provided for traffic management, they do not do enough to achieve smooth running of traffic in large cities. Even in the case of more congested roads where traffic is more, fog computing nodes also faces some issues such as load distribution, more bandwidth consumption, and more energy consumption in nodes (Ning Z. a., 2019).