Distributed Performance Evaluation of 5G and Wi-Fi for Private Industrial Networks

Wireless communication plays an increasingly large role in modern industrial environments, enabling highly flexible manufacturing processes. While Wi-Fi is still a dominant technology, private 5G networks promise to provide a sophisticated, more reliable solution in licensed frequency bands. However, it remains uncertain in practice which level of end-to-end performance can actually be achieved. To this end, this contribution focuses on a technology-independent performance evaluation with distributed test devices based on the STING concept (Spatially Distributed Traffic and Interference Generation). We present a comparative analysis of 5G and $\mathrm{Wi}-\mathrm{Fi}$ technologies for private industrial networks under realistic network load scenarios, focusing on latency to determine their suitability within industrial contexts. It can be seen that the current Wi-Fi 6 technology can keep critical applications reliable in low to medium load scenarios, but struggles when multiple stations have critical Quality of Service (QoS) requirements. In contrast, it is shown that 5G technology maintains lower latency times and has fewer outlier effects that are considered unacceptable in functional safety applications, especially in scenarios with medium to high network utilization. In order to maintain the advantage of simple Wi-Fi technology while increasing latency stability, we have optimized channel access (EDCA) in a further step on the basis of an open source Wi-Fi stack. It is shown that Wi-Fi technology with suitable modifications beyond the standard can also be able to reliably serve several stations with high $\mathbf{Q o S}$ requirements.