Better Safe Than Sorry: Distributed Testbed for Performance Evaluation of Private Networks

Abstract

The operation of local private cellular radio networks in licensed frequency bands constitutes one of the core innovations of current 5G and future 6G networks. The predicted performance of an exclusively utilizable and thus interference-free frequency range of, for example, private 5G networks, so-called campus networks, are generally of great interest to industrial companies. However, the integration of 5G network infrastructures into existing brownfield environments has to overcome major technological and administrative challenges. In contrast to very rarely encountered greenfield scenarios, the potential of 5G can only be realized in practice if significant performance advantages over existing wireless network infrastructures (e.g., Wi-Fi) can be demonstrated while guaranteeing seamless integration into the process landscape. To this end, this paper presents an agile system for continuous, cross-network monitoring of end-to-end guarantees in terms of throughput, latency and reliability. While single pointwise measurements during network deployment often indicate the expected performance peaks, this contribution specifically explores the potential of a spatially distributed stress test that actively monitors network quality on a continuous basis. An extensive case study is conducted to demonstrate the performance of the distributed approach for performance evaluation of multi-user and cell edge environments. In addition, it is illustrated that the distributed system can be used to estimate how mission-critical service guarantees affect the overall network performance.