An event-driven and lightweight proactive auto-scaling architecture for cloud applications

Abstract

The cloud environment is used by the application providers (APs) to host their applications in order to reduce procurement and management costs of the cloud resources. Moreover, the variation in the traffic load of the client applications and the appealing auto-scaling capability of the cloud resources have prompted application providers to seek ways to reduce the cost of their rented services. This paper describes a constructive auto-scaling mechanism based on the events in cloud systems fitted with heuristic predictors. The predictor examines historical data using these approaches: (1) Double Exponential Smoothing (DES), (2) Triple Exponential Smoothing (TES), (3) Weighted Moving Average (WMA) and (4) WMA with Fibonacci numbers. The outcomes of this model simulation in CloudSim indicate that the model can decrease the application provider's cost while preserving application user satisfaction.