Volume - 6 | Issue - 4 | december 2024
Prediction of Workloads in Cloud using ARIMA-ANN
Pages: 327-342
Cite this article
S., Suriya, and Surya Arvindh M.. "Prediction of Workloads in Cloud using ARIMA-ANN." Journal of IoT in Social, Mobile, Analytics, and Cloud 6, no. 4 (2024): 327-342
Published
16 January, 2025
Abstract
Full Article PDF
This study introduces an innovative hybrid ARIMA-ANN model personalized for cloud workload prediction. Unlike existing models that focus solely on linear or nonlinear patterns, the approach combines the strengths of ARIMA for time-series linear trends and ANN for nonlinear data complexities. This integration ensures higher accuracy, as validated using the MIT Supercloud dataset. The methodology leverages data pre-processing, sensitivity analysis, and advanced validation techniques, demonstrating improved accuracy in scenarios of high workload variability. This model supports cloud providers in resource optimization and dynamic load management.
KeywordsARIMA ANN Load Balancing Time Series data MIT Supercloud
