Abstract

Narrow Band Internet of Things (NB-IoT) is a promising radio technology that was standardized by 3GPP in 2016 for connecting a massive number of low-cost, low-power and delay-tolerant devices to the Internet of Things (IoT) with a small bandwidth of 180 KHz. Prior studies on the random access of NB-IoT have investigated its performance under various conditions and considering several mechanisms, however, the backoff mechanism as a method for contention resolution was ignored mostly either for simplification or in favor of other mechanisms while few works considered it under certain restrictions. This thesis proposes a comprehensive analytical model that evaluates the performance of random access procedures under the assumption of adaptive backoff mechanism that doubles the backoff window size for each new transmission attempt to achieve a better contention resolution.
The proposed model allows arrival of packets with different rates to each device in the network which is more practical in cases where there are noticeable differences in specifications between connected devices such as in layered, packet loss intolerant and clustered networks. The system is modeled using discrete-time analysis with First-Come-First-Served (FCFS) user queues with infinite buffers. The analysis has led to derivations of probability of successful packet transmission, mean packet delay, utilization, and probability of packet discarding. The results have shown the advantages of the adaptive backoff mechanism in improving the performance of the system compared to constant backoff. Additionally, results have demonstrated the trade-off between packet discarding probability and mean packet delay where higher number of attempts allows for almost zero packet loss probability at the cost of slightly higher delay. A simulation system was implemented which verified the accuracy of the analytical model. The results of this thesis can be helpful in designing new NB-IoT communication networks.