Evaluating Duty-Cycle Enforcement in Distance-Vector LoRa Mesh Networks: A Cross-Layer ns-3 Study
Multi-hop LoRa (LoRa mesh) can extend coverage and resilience beyond LoRaWAN star topologies, but its end-to-end performance degrades when duty-cycle (DC) constraints are enforced and traffic is no longer sporadic. However, many LoRa mesh evaluations either do not enforce DC explicitly or do not record metrics that connect DC enforcement with control-plane degradation. This paper evaluates the impact of DC enforcement in a cross-layer LoRa mesh stack with proactive Distance-Vector (DV) routing that advertises routes through a compact scalar score and a CSMA/CA MAC based on Channel Activity Detection (CAD), with explicit 1% DC applied to both data and beacons. We evaluate this stack under any-to-any traffic in a $1 ~\text{km}^{2}$ area for network sizes $N=9,16,25,36,49$, and 64 randomly located nodes, under low and medium loads, and 10 seeds per configuration (240 runs). Results show that enabling DC systematically reduces the end-to-end delivery ratio (DR): for $N=25$ it drops from 0.958 to 0.438 under low load and from 0.159 to 0.058 under medium load, while for $N=64$ the relative impact is smaller (from 0.399 to 0.314 and from 0.051 to 0.039) due to a transition to a contention-dominated regime. Control-plane logging is consistent with this mechanism: DC reduces beacon health $\left(H_{b}\right)$, which is accompanied by DV route expirations and more no-route events, linking DC enforcement to end-to-end losses.