RNIS-Aware Priority Flow Protection in Softwarized 5G NTN-MEC

Kurzfassung

Low Earth Orbit (LEO) satellite networks introduce time-varying delay, intermittent connectivity, and dynamic link conditions that can degrade Quality of Service (QoS) in 5G Non-Terrestrial Networks (NTN). This paper proposes an RNIS aware traffic-shaping framework for priority flow protection in a softwarized NTN-MEC system. By using Radio Access Network (RAN)-side Key Performance Indicator (KPI) exposure through the Radio Network Information Service (RNIS), the proposed approach enables adaptive IP layer traffic control before traffic enters the constrained NTN-RAN segment. The framework is implemented on an OpenAirInterface (OAI)-based NTN-MEC testbed with a system-level emulated LEO channel. A window based control mechanism combines normalized RNIS-side in dicators with end-to-end measurements to regulate competing traffic. Experimental results show that the evaluated approach improves priority flow performance under the same offered-load conditions, increasing priority flow throughput from 1.033 Mbit/s to 1.669 Mbit/s while reducing RTT-p95 from 26.30 s to 6.94 s compared with the unshaped case. These results demonstrate that RNIS-assisted pre-bottleneck traffic shaping can mitigate congestion effects and improve QoS protection for priority traffic under the evaluated emulated NTN conditions.