Impact of Seasonal ARIMA workload prediction model on QoE for Massively Multiplayers Online Gaming

Abstract

Ensuring an acceptable Quality of Experience (QoE) for all users is a fundamental requirement to the economical development of the Massively Multiplayers Online Gaming (MMOG) companies. However, the high load variability of such MMOG services makes hard to satisfy a good QoE. This paper aims to contribute to this effort, by proposing a proactive dynamic provisioning approach which predicts future workload of an MMOG service and allocates in accordance the sufficient amount of resources. Based on real MMOG traces, we propose a Seasonal Autoregressive Integrated Moving Average (SARIMA) model that generally fits the workload behavior of the MMOG cloud service. We implement our prediction-based algorithm that allocates resources according to predicted workload by SARIMA model. Finally, we evaluate impact of our proposed algorithm on the QoE, where experiments prove noticeable improvements.