Detection and Replay of Distributed Denial of Service Attacks In Smart Cities Using a Hybrid Deep Learning Approach

Abstract

Infrastructure for smart cities is presently being built, mostly because of the (IoT) platform, which connects a wide variety of things by using the website's underlying infrastructure. As a result, it is possible to employ a uniform platform to automate the services that are delivered. Smart city infrastructure, on the other hand, is susceptible to cyberattacks owing to security flaws in IoT networks. Distributing denial of service (DDoS) and replay assaults, for example, violate the requirements for certification in smart cities. The lack of citations in this part backs up many of the statements it makes. Who needs a citation? Both assaults have the potential to substantially damage smart city infrastructure, which might also cost money and even result in human fatalities. This paper covers the creation of a blended deep- learning algorithm for replay and DDoS intrusion prevention on a growth-oriented smart city platform. This is made possible by combining proven machine-learning approaches with more recent advances in artificial intelligence. We simulate scattered denial of service and replay attacks on three datasets acquired from real-world smart cities to assess the efficacy of the proposed hybrid strategy on environmental, smart river, and smart soil dataset. The suggested model has shown excellent rates of accuracy. The ecosystem collection has an accuracy level of 98.37%, the smart riverbed dataset must have an accuracy of 98.13%, and the smart mud dataset must have an accuracy of 99.51%. The findings demonstrated that the proposed model outperformed past instances of evolutionary computation and machine learning techniques employed in the corpus of academic literature.