Minimized False Alarm Predictive Threshold for Cloud Service Providers

Title:Minimized False Alarm Predictive Threshold for Cloud Service Providers

Volume: 14 Issue: 5

Author(s): Amandeep S. Arora , Linesh Raja *Barkha Bahl

Affiliation:

• Department of Computer Applications, Manipal University Jaipur, Rajasthan,India

Keywords: Machine learning, DDOS attack, cloud computing, cross validation, classification, feature selection.

Abstract:

Aim: Cloud Security is a strong hindrance which discourages organisations to move toward cloud despite huge benefits. Distributed denial of service attacks operated via distributed systems compromise availability of cloud services which cause limited resources for authentic users and high expense for cloud service users and business owners.

Objective: Techniques to identify distributed denial of service attacks with minimized false positives are highly required to ensure availability of cloud services to genuine users. Scarcity of solution which can detect denial of service attacks with minimum false positives and reduced detection delay has motivated us to compare classification algorithms for detection of distributed denial of service attacks with minimum false positive rate.

Methods: Classification of incoming requests and outgoing responses using machine learning algorithms is a quite effective way of detection and prevention. We designed a performance tuned support vector machine algorithm with features of F-hold cross validation strategy.

Results: F-hold crosses validation strategy, which can detect denial of service packets with 99.89% accuracy.

Conclusion: This system ensures economic sustainability for business owners and limited resources mitigation for authenticated and valid cloud users.

Cite this article as:

Arora S. Amandeep , Raja Linesh*, Bahl Barkha , Minimized False Alarm Predictive Threshold for Cloud Service Providers, Recent Advances in Computer Science and Communications 2021; 14 (5) . https://dx.doi.org/10.2174/2213275912666190819115320