RHAS: robust hybrid auto-scaling for web applications in cloud computing

Abstract

The elasticity characteristic of cloud services attracts application providers to deploy applications in a cloud environment. The scalability feature of cloud computing gives the facility to application providers to dynamically provision the computing power and storage capacity from cloud data centers. The consolidation of services to few active servers can enhance the service sustainability and reduce the operational cost. The state-of-art algorithms mostly focus either on reactive or proactive auto-scaling techniques. In this article, a Robust Hybrid Auto-Scaler (RHAS) is presented for web applications. The time series forecasting model has been used to predict the future incoming workload. The reactive approach is used to deal with the current resource requirement. The proposed auto-scaling technique is designed with the threshold-based rules and queuing model. The security mechanism is used to secure the user’s request and response to the web-applications deployed in cloud environment. The designed approach has been tested with two real-time web application workloads of ClarkNet and NASA. The proposed technique achieves $$14\%$$ reduction in cost, and significant improvement in response time, service level agreement (SLA) violation, and gives consistency in CPU utilization.