Modeling Correlation between QoS Attributes for Trust Computation in Cloud Computing Environments
Manel Mrabet, Yosra Ben Saied, Leı̈la Azouz Saı̈dane · 2017
This paper introduces an approach that handleswith the trustworthy cloud service selection issue in Cloudcomputing environments. Despite the fact that most of theexisting trust systems consider several QoS attributes for trustcomputing, none of them did consider the correlation that mayexist among these attributes. However, we demonstrate in thispaper that the integration of correlation between QoS attributesin trust computing significantly helps to solve many issues suchas predicting missing assessment, detecting malicious feedbacksand improving the accuracy of trust values. The main goal ofthis paper is to show the significance of correlation betweenthose attributes for trust computing. To do that, we proposecombining both the popular Naïve Bayes model and the ngramMarkov model to design a more efficient trust model forcloud environments. Our proposed trust model also takes intoaccount the user's requirements, aggregates both the qualitativeand quantitative assessment, and considers several sources whencomputing cloud services' trust values. Experimental results showthat our proposed approach outperforms the traditional Naïve Bayes trust models.