<  Back to the Polytechnique Montréal portal

An SVM-based framework for detecting DoS attacks in virtualized clouds under changing environment

Adel Abusitta, Martine Bellaïche and Michel R. Dagenais

Article (2018)

[img]
Preview
Published Version
Terms of Use: Creative Commons Attribution.
Download (1MB)
Cite this document: Abusitta, A., Bellaïche, M. & Dagenais, M. R. (2018). An SVM-based framework for detecting DoS attacks in virtualized clouds under changing environment. Journal of Cloud Computing : Advances, Systems and Applications, 7(9). doi:10.1186/s13677-018-0109-4
Show abstract Hide abstract

Abstract

Cloud Computing enables providers to rent out space on their virtual and physical infrastructures. Denial of Service (DoS) attacks threaten the ability of the cloud to respond to clients requests, which results in considerable economic losses. The existing detection approaches are still not mature enough to satisfy a cloud-based detection systems requirements since they overlook the changing/dynamic environment, that characterises the cloud as a result of its inherent characteristics. Indeed, the patterns extracted and used by the existing detection models to identify attacks, are limited to the current VMs infrastructure but do not necessarily hold after performing new adjustments according to the pay-as-you-go business model. Therefore, the accuracy of detection will be negatively affected. Motivated by this fact, we present a new approach for detecting DoS attacks in a virtualized cloud under changing environment. The proposed model enables monitoring and quantifying the effect of resources adjustments on the collected data. This helps filter out the effect of adjustments from the collected data and thus enhance the detection accuracy in dynamic environments. Our solution correlates as well VMs application metrics with the actual resources load, which enables the hypervisor to distinguish between benignant high load and DoS attacks. It helps also the hypervisor identify the compromised VMs that try to needlessly consume more resources. Experimental results show that our model is able to enhance the detection accuracy under changing environments.

Uncontrolled Keywords

Cloud computing, DoS attacks detection, Support vector machine, Changing environment, Virtual machines

Open Access document in PolyPublie
Subjects: 2700 Technologie de l'information > 2716 Réalité virtuelle et simulations connexes
2700 Technologie de l'information > 2719 Architecture d'ordinateur et conception
2700 Technologie de l'information > 2721 Systèmes et réseaux multimédias
Department: Département de génie informatique et génie logiciel
Research Center: Non applicable
Funders: CNSNG/NSERC
Date Deposited: 07 Dec 2020 11:28
Last Modified: 08 Apr 2021 10:42
PolyPublie URL: https://publications.polymtl.ca/4735/
Document issued by the official publisher
Journal Title: Journal of Cloud Computing : Advances, Systems and Applications (vol. 7, no. 9)
Publisher: Springer
Official URL: https://doi.org/10.1186/s13677-018-0109-4

Statistics

Total downloads

Downloads per month in the last year

Origin of downloads

Dimensions

Repository Staff Only