A Real-Hardware Multi-Instance Evaluation of ETSI DCC

Decentralized Congestion Control (DCC) is a key mechanism for guaranteeing scalable and reliable Vehicle-to-Everything (V2X) communications, especially in dense traffic scenarios. Despite several analytical and simulation-based studies, extensive experimental evaluations with real V2X hardware of complete standard-compliant DCC implementations remain limited. This paper presents an experimental assessment of DCC based on a fully standard-compliant, open-source implementation integrated into the OScar V2X software stack. The proposed system implements the complete DCC architecture, including Access-Layer (DCC ACC), Network-Layer (DCC NET), and Cross-Layer (DCC CROSS) components, and is evaluated using a controlled laboratory testbed composed of nine On-Board Units (OBUs), each carrying three instances of the protocol stack, emulating a realistic vehicular scenario with vehicles traveling as a platoon. Mobility and perception dynamics are faithfully reproduced through the replay of traces collected during real-world highway driving experiments. Experimental results under varying network load highlight the effectiveness of adaptive and reactive congestion control algorithms in maintaining channel stability and reliable message dissemination, while also revealing practical performance trade-offs in dense vehicular environments.