Realistic Traffic Generation Orchestrator for 5G Networks

The evolution of Fifth Generation (5G) networks demands experimental platforms that can replicate the stochastic nature of real-world traffic. However, traditional benchmarking tools often rely on stationary models that fail to capture the burst behavior of modern applications. This paper presents a novel orchestration framework for generating realistic, non-stationary traffic for 5G testbeds. The solution automates heterogeneous application profiles, including interactive browsing, Dynamic Adaptive Streaming over HTTP (MPEG-DASH) streaming, and Cloud gaming simulation across multiple User Equipments (UEs) without human intervention. Experimental results demonstrate that the orchestrator effectively mimics unpredictable load fluctuations, providing a high-fidelity environment for network performance evaluation that significantly surpasses conventional synthetic traffic models.