Memory-effect Based QoE Evaluation Method and Guarantee Scheme in APN-driven Game Acceleration

Application-aware Networking (APN) is crucial for ensuring Quality of Experience (QoE) in real-time online games. The adversarial nature of real-time online games introduces a memory-effect in network fluctuations, where the accumulation of past QoE of players over time may influence their current QoE. This phenomenon adds complexity to QoE evaluation and impacts the APN-driven game QoE guarantee.To address this,we propose MEEG, a scheme aimed at evaluating and guaranteeing the QoE of real-time online games from the perspective of the memory-effect. MEEG introduces a memory-effect based QoE evaluation method, capturing the existence, time dependency, and intensity of the memory-effect. Then, an algorithm was developed to tackle the APN-driven game QoE guarantee problem, considering player QoE requirements and aiming to minimize resource costs in determining APN parameters. MEEG is applied to the popular real-time online game, Arena Of Valor, to create its specific QoE model. Experimental verification involving ten volunteers demonstrates that MEEG achieves a more precise QoE evaluation in network fluctuations, reducing the average mean square error(MSE)byupto66.42%compared to existing schemes. Additionally, MEEG's QoE guarantee ability to withstand delay fluctuations increases from 150ms to 350ms while reducing resource consumption by up to 40%.