LoRaWAN is one of the most commonly used Internet of Things protocols for applications that require low cost, low power, and long range communications. It has been proved that – mainly due to regional radio duty cycle restrictions – the protocol does not scale well in presence of confirmed (downlink) traffic. To support downlink traffic, LoRaWAN employs two receive windows, RX1 and RX2, whereas a number of channels are assigned to each of those windows. The protocol uses a fixed Spreading Factor (SF) – a LoRa PHY modulation parameter – in RX2, while the SF of the uplink is employed in RX1. Since the SF of RX1 cannot be changed, selecting a low or a high value of SF in RX2 is of critical importance for the duty cycle resources of the gateways. On one hand, selecting high SF values, the time resources of the gateways may get depleted fast leading to low capacity because the transmission time increases with higher SF values. On the other hand, lower SF values reduce reachability due to the worse sensitivity which causes retransmissions, and thus, lower capacity. In this paper, a study of the total theoretical downlink capacity is provided giving useful insights of the protocol behavior as the number of uplinks increases, especially for congested network scenarios. An exhaustive SF selection solution is also presented to compute the maximum downlink capacity. It is shown that the RX2 SF which provides the best capacity may imply fairness compromises for part of the devices. Extensive simulations are employed to confirm the theoretical findings in scenarios with a single and multiple gateways. Experiments conducted on a test-bed for selected scenarios also confirm the theoretical findings.
Read full abstract