Deep Learning-based Workload Prediction in Cloud Computing to Enhance the Performance

The cloud paradigm offers a plethora of valuable services to end users through virtualization, aiming to improve resource utilization and reduce services cost. But unpredictable demand of end users degrades the service performance and loss of revenues. The prediction of mixed and inseparable workloads in cloud computing plays a vital role to allocate resources efficiently. Hence, an efficient prediction model is required to give better results in workload management, job scheduling, and overall efficiency of cloud servers. This paper proposes a deep learning-based approach that uses a combination of neural networks to analyze historical workload data and forecast future workloads. The model is evaluated using real existing datasets which are verified and collected by the real-time problems and the output of the proposed algorithm accurately predicts future workloads and improves the user satisfaction level and system efficiency.