An Intrusion Detection System Using Modified-Firefly Algorithm in Cloud Environment

An Intrusion Detection System Using Modified-Firefly Algorithm in Cloud Environment

Partha Ghosh (Netaji Subhash Engineering College, India), Dipankar Sarkar (Tata Consultancy Service, India), Joy Sharma (Netaji Subhash Engineering College, India) and Santanu Phadikar (Maulana Abul Kalam Azad University of Technology, India)
Copyright: © 2021 |Pages: 17
DOI: 10.4018/IJDCF.2021030105
Article PDF Download
Open access articles are freely available for download

Abstract

The present era is being dominated by cloud computing technology which provides services to the users as per demand over the internet. Satisfying the needs of huge people makes the technology prone to activities which come up as a threat. Intrusion detection system (IDS) is an effective method of providing data security to the information stored in the cloud which works by analyzing the network traffic and informs in case of any malicious activities. In order to control high amount of data stored in cloud, data is stored as per relevance leading to distributed computing. To remove redundant data, the authors have implemented data mining process such as feature selection which is used to generate an optimum subset of features from a dataset. In this paper, the proposed IDS provides security working upon the idea of feature selection. The authors have prepared a modified-firefly algorithm which acts as a proficient feature selection method and enables the NSL-KDD dataset to consume less storage space by reducing dimensions as well as less training time with greater classification accuracy.
Article Preview
Top

Introduction

Cloud Computing is one of the upcoming technologies which provides software, platforms and infrastructural services as per the requirement of users (Yeboah-Boateng & Essandoh, 2013). Cloud provides service to the people to store and make use of stored materials for their purpose. There is a universal access to the user’s data throughout the world using any Internet-ready devices (M. Davis & A. Sedsman, 2010). Cloud computing is basically associated with the Distributed computing. Distributed computing is required in situation where the data is so huge that it cannot be saved in a single storage device, rather data is stored in distributed manner in Cloud Environment. Although with the increasing acceptance of Cloud, the flow of attacks commonly called intrusion to the system is also increasing (Akbar, Dr.K.Nageswara, & Dr.J.A.Chandulal, 2010). Data saved in the Cloud are highly sensitive and important, so data security is a major concern to protect them from intruder. The malicious attacks affect the properties of storage system such as confidentiality, integrity and availability (A.E. Azzouzi & K.E.E. Kadiri, 2015). IDS is an approach to provide security and shorten the damage of stored data. It provides software and hardware services to put a check on the security, analyses for malicious activity and also produces a report to the management system (Araújo & Abdelouahab, 2012). To detect any kinds of abuse or crime using network, which does not obey the law, IDS is required to build. Network forensic is a method of capturing, storing and analyzing data to find the source of security violator. To detect security violator two types of IDS is there. Host based Intrusion Detection System(HIDS) runs on a particular hosts or machines on the network. HIDS supervises the incoming and outgoing traffic from the machines only and will notify the user or administrator if any unusual action is detected (Partha Ghosh, Ghosh, & Dutta, 2014). Network based Intrusion Detection System(NIDS) are deployed in the crucial points. NIDS watches for abuse of protocols, curious patterns and supervises user actions. It works on the traffic and detects network anomalies (Agrawal & Kamble, 2012). The Network Design should be such that the IDS classifies all types of connections into both normal and abnormal. Artificial Neural Networks can improve the performance and efficiency of IDS (C. Lu, Y. Li, M. Ma & N. Li, 2016). Data mining is the method of exploring patterns from huge data sets involving methods by the combination of machine learning and statistics. As this involves a very large training dataset that increases memory space requirement as well as time. Therefore to reduce the number of features a Feature Selection(FS) algorithm is required which is a Data mining process (P. Ghosh & Mitra, 2015). In this paper, the authors have approached the feature selection using the proposed Modified-Firefly Algorithm (MFA) to reduce the dimension of the dataset. Using the datamining process, after analyzing the network activity and comparing with benchmark signatures network forensics and crimes can be detected. To make the analyzing and detecting process faster and more accurate the author proposed MFA. This Modified Firefly Algorithm is achieved by blending the idea of three different data mining methods. These methods includes PSO, FA and Fuzzy Logics. By exhibiting the results of the experiment it can be concluded that the proposed MFA produces better optimum feature subset than FA. Finally, a number of classifiers have been used namely AdaBoost, Neural Network and Random Forest on this reduced dataset obtained by applying MFA to find the accuracy of the proposed IDS Model. The outcome of these classifiers exhibits improved accuracy.

Complete Article List

Search this Journal:
Reset
Volume 14: 1 Issue (2022)
Volume 13: 6 Issues (2021)
Volume 12: 4 Issues (2020)
Volume 11: 4 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing