JISE


  [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]


Journal of Information Science and Engineering, Vol. 36 No. 5, pp. 981-992


Adaptive Learning based Prediction Framework for Cloud Datacenter Networks' Workload Anticipation


JITENDRA KUMAR1,2 AND ASHUTOSH KUMAR SINGH2
1Department of Computer Applications
National Institute of Technology Tiruchirappalli
Tamilnadu, 620015 India

2Department of Computer Applications
National Institute of Technology Kurukshetra
Haryana, 136119 India
E-mail: jitendrakumar@ieee.org; ashutosh@nitkkr.ac.in


Cloud computing has effectively changed the computing industry by introducing the on-demand resources through virtualization. However, a cloud system suffers with several challenges including low resource utilization, high power consumption, security and many others. This paper introduces a neural network based workload forecasting model using differential evolution. The predictive framework is evaluated on five real world data traces. The forecast efficacy is compared with state-of-art approaches including back propagation and linear regression along with statistical analysis. It was observed that the proposed scheme reduced the forecast error up to 85.52% and 89.70% measured using RMSE and MAE respectively. The statistical analysis also validates the superiority of the proposed predictive framework as it received the best rank in the Friedman test analysis.


Keywords: workload forecasting, differential evolution, neural network, cloud computing, Google cluster trace

  Retrieve PDF document (JISE_202005_03.pdf)