IIARD International Institute of Academic Research and Development

INTERNATIONAL JOURNAL OF APPLIED SCIENCES AND MATHEMATICAL THEORY (IJASMT )

E- ISSN 2489-009X
P- ISSN 2695-1908
VOL. 9 NO. 2 2023
DOI: https://doi.org/10.56201/ijasmt.v9.no2.2023.pg66.83

---

Investigating Cloud Service Attacks using Machine Learning

Durunna Lilian Ijeoma., Agbakwuru A.O., Agbasonu V.C., Amanze B.C

---

Abstract

In this work, a rigorous investigation has been carried out as to the classification of cloud service attacks, utilizing machine learning algorithms. Examining the attacks emphasized that those features in the dataset are recognise as the most significant when it comes to identifying the cloud service attacks. When seeking to establish identification, cloud service attacks are viewed as being the most troublesome, predominantly considering their involvement of network- and host-level characteristics. As such, both host and network-level components-namely ‘duration of connection’ and ‘service requested’, and the ‘number of failed login attempts’, respectively-are selected in the establishment of cloud service attacks. In considering the form of operation manipulated by the cloud service attacks, the significant features when it comes to recognizing such forms are seen to belong to the KDD cup dataset. Importantly, most cloud applications present the potential of various logins, which fundamentally positions the end user in such a way that they are able to link to the email server through various instruments simultaneously. As such, traffic features need to be determined through the presence of a host–services connection incorporating a two-second time window. In line with the Ftp exploitation tool, which is generalized as an attack on the FTP protocol through the attacker making use of the PORT command with the aim of achieving access to ports, this is commonly recognise as an Ftp bounce attack. Accordingly, in Ftp identification, the aspects of urgent, compromised, -access-files and host-count are essential, as determined through the methods as being unique features for Ftp. In regards to the password guessing form

keywords:

Machine Learning, attacks, cloud computing platform, Financial Institution

---

References:

Dunlop, M.; Groat, S.; Shelly, D. (2020) GoldPhish: Using Images for Content-Based Phishing
Analysis, Internet Monitoring and Protection (ICIMP), 2020 Fifth International
Conference, pp.123-128, May 2020.

Jun, F.; Yu, C.; Wei-Shinn, K.; Pu, L. (2019) Analysis of Integrity Vulnerabilities and a Non-
repudiation Protocol for Cloud Data Storage Platforms, Parallel Processing Workshops
(ICPPW), 2019 39th International Conference, pp.251-258, Sept. 2019.

DOWNLOAD PDF

---

Back