Home // CLOUD COMPUTING 2025, The Sixteenth International Conference on Cloud Computing, GRIDs, and Virtualization // View article

Proactive Optimization of Virtual Machine Placement Using Predictive Models Based on Time Series

Authors:
Naby Doumbouya
Mhand Hifi

Keywords: Cloud computing, virtual machine placement, time sequences, ARIMA, LSTM, Prophet, cloudsim,Time series fore- casting,Proactive optimization

Abstract:
The rapid emergence of cloud computing has significantly increased the need for effective virtual machine (VM) placement optimization strategies. The primary goal of these strategies is to maximize energy efficiency, reduce operational costs, and ensure optimal quality of service through intelligent resource management. However, traditional approaches based on reactive scheduling have proven insufficient in the face of dynamic workload variations, often resulting in excessive energy consumption, performance degradation, and unnecessary VM migrations, which are particularly costly in terms of both resources and time. To address these limitations, we propose in this paper an innovative proactive strategy based on predictive time series analysis, enabling accurate anticipation of future VM workload variations. We rigorously evaluate and compare three major predictive models: the statistical ARIMA model, the Long Short- Term Memory (LSTM) recurrent neural network, and the Prophet model, specifically designed to handle time series data. This evaluation focuses on the prediction accuracy of these models and their ability to significantly reduce unnecessary VM migrations while improving overall system performance. Finally, the predictions are integrated into an advanced optimization algorithm to proactively determine the optimal placement of VMs before workload spikes occur. Our approach thus demonstrates its effectiveness through better anticipation of workload fluctuations, a notable reduction in energy costs, and a substantial improvement in stability and quality of service within distributed data centers.

Pages: 122 to 124

Copyright: Copyright (c) IARIA, 2025

Publication date: April 6, 2025

Published in: conference

ISSN: 2308-4294

ISBN: 978-1-68558-258-6

Location: Valencia, Spain

Dates: from April 6, 2025 to April 10, 2025