Dynamic QoS-Aware Cloud Service Selection Using Best-Worst Method and Timeslot Weighted Satisfaction Scores

The number of cloud services has dramatically increased over the past few years. Consequently, finding a service with the most suitable quality of service (QoS) criteria matching the user's requirements is becoming a challenging task. Although various decision-making methods have been proposed to help users to find their required cloud services, some uncertainties such as dynamic QoS variations hamper the users from employing such methods. Additionally, the current approaches use either static or average QoS values for cloud service selection and do not consider dynamic QoS variations. In this paper, we overcome this drawback by developing a broker-based approach for cloud service selection. In this approach, we use recently monitored QoS values to find a timeslot weighted satisfaction score that represents how well a service satisfies the user's QoS requirements. The timeslot weighted satisfaction score is then used in Best-Worst Method, which is a multi-criteria decision-making method, to rank the available cloud services. The proposed approach is validated using Amazon's Elastic Compute Cloud (EC2) cloud services performance data. The results show that the proposed approach leads to the selection of more suitable cloud services and is also efficient in terms of performance compared to the existing analytic hierarchy process-based cloud service selection approaches.