ABSTRACT
LoRaWAN (long range wide area network) is an emerging wireless communication network that can provide low power transmissions and wide area coverage for IoT (internet of things) end devices. Since it operates in an ISM (industrial scientific medical) band, there is a regulation of duty cycle, e.g., 1%, for each sub-band in uplink transmissions. In addition, there may be several spreading factors available for an end device to select for uplink transmissions, and the service rate of using one spreading factor is different than that of using another. In literature, for simplicity, a uniform service rate is assumed, so that the uplink transmissions are modeled as a M/D/c queueing system. Though an approximated mean queuing delay can be obtained from the simplified model, the accuracy is far from satisfactory. In this paper, a Markov process is first utilized to reflect different service rates between spreading factors, and the uplink transmissions of an end device can then be modeled as a M/G/c queueing system. Numerical results show that more accurate mean queueing delay can be obtained under our proposed analysis.
