International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 07 | July 2023
p-ISSN: 2395-0072
www.irjet.net
Deep Learning and Big Data technologies for IoT Security Pranesh Kathavate1, Suvidhi Solanki2 1Student, Computer Engineering Dept, Vishwakarma University, Pune, India 2Student, Computer Engineering Dept, Vishwakarma University, Pune, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - An As there are millions to billions of
potential to gather, quantify, and grasp data about their surroundings, enabling modernizations that improve people's quality of life[3].The Internet of Things will encompass all aspects of our lives, with applications ranging from home automation to smart transportation to smart agriculture to wearable devices and e-Health. IoT devices are easy targets for botnet developers due to their vulnerability. The proliferation of IoT malware presents new security challenges for network managers. In this part, we will first describe the intrinsic characteristics of IoT ecosystems that make them difficult to safeguard. Then, we discuss the most frequent types of vulnerabilities found in IoT ecosystems, as well as how IoT botnets operate. We then explain a number of proposed solutions for IoT security, as well as their drawbacks. Deep learning techniques are widely used for high-dimensional object analysis. This is owing to their high ability to generalize data. The method for adjusting such structures is focused at approximating the formation of such a set that includes both training elements and elements not encountered during the training process. In [4], the author suggested a deep learning-based approach for detecting internet intrusions in the IoT. The authors underline the framework's effectiveness within the context of the ''smart city'' notion. The investigated strategy has the advantage of improving the classification accuracy and creating a trade-off balance between the correctness of attack detection and the speed of this process. In [5}, the research focuses on developing a distributed attack detection system using a deep learning model. The created system's application is the detection of assaults in the social IoT.
connected Internet of Things (IoT) devices and systems sending enormous raw and processed data through the IoT network, we need to be able to use big data analytical techniques and solutions with an effective classification of possible attacks using deep learning in order to protect the security and privacy of IoT data and services from a wide range of attackers. The vast volume of organized, semistructured, and unstructured data that is produced in the digital era has made the term "big data" popular in recent years. But this vast amount of data is trackable and can be used for financial gain, which violates people's privacy. The attack surfaces of the IoT system are investigated, along with any potential risks related to each surface. The use of deep learning to find security flaws has proved successful in earlier research. The data generated by IoT devices is abundant, diverse, and reliable. We then discuss the advantages, disadvantages, and strengths of each Deep Learning technique for IoT security. As a result, big data technology makes it possible to organize and perform tasks better. In this paper, we have focused on combining the potentiality of Big Data processing with Deep Learning Key Words: IoT Security, Big Data, Deep Learning, Botnet, DDoS, Access Control, Malware detection, RNN
1.INTRODUCTION The number of common devices equipped with sensors and capable of internet communication has significantly increased during the last few years. According to an article in IoT Business News, the number of linked devices in the IoT globe is predicted to reach about 75 billion by 2030.[1] These devices are capable of recognizing their current situation, sharing, and managing data that can be used for a variety of purposes. The Internet of Things (IoT) was quickly developed and adopted in a variety of industries, including business, agriculture, and the military. Big data is getting bigger and more complicated, especially from new sources. These enormous datasets can't be processed using conventional methods. Despite the overwhelming amount of data, it might be helpful in resolving business issues that you were previously unable to handle. These gadgets are equipped with a range of sensors that enable them to collect data in real time from distant physical devices. For example, the Internet of Things (IoT) has substantially increased traditional detection of surrounding situations. IoT technologies have the
© 2023, IRJET
|
Impact Factor value: 8.226
1.1 Big Data Big data is being particularly collected in massive amounts, which raises complexity. Modern technology is unable to handle these large data quantities. Innovative data processing techniques are necessary because of the massive volume of data collected and its quick change. This terminology is required to improve judgment, gain a deeper understanding of processes, and make them function more effectively. We may refer to data as "Big Data" if traditional or modern technology has trouble gathering, processing, storing, filtering, and visualizing it. In a word, "Big Data" technology involves gathering, storing, and extracting knowledge from enormous data volumes. Huge data is a term for a significant volume of
|
ISO 9001:2008 Certified Journal
|
Page 896