A massive number of internet-of-things (IoT) and machine-to-machine (M2M) communication devices generate various types of data traffic in cellular IoT networks: periodic or non-periodic, bursty or sporadic, etc. In particular, bursty and non-periodic traffic may cause an unexpected network congestion and temporary lack of radio resources. In order to effectively accommodate such bursty and non-periodic traffic, we propose a novel recursive access class barring (R-ACB) technique to optimally utilize the available resources associated with the random access procedure (RAP) which consists of multiple steps in cellular IoT networks, while existing ACB schemes only considered the resource of the first step of RAP, i.e., the number of available preambles. The proposed R-ACB technique consists of two main parts: 1) online estimation of the number of active IoT/M2M devices who have data to transmit to an eNodeB and 2) adjustment of the ACB factor which indicates the probability that an active device sends a preamble to the eNodeB. It is notable that the estimation and the adjustment recursively affect each other when R-ACB operates. In addition, we also propose mathematical models to analyze the performance of R-ACB in terms of total service time, average access delay, resource efficiency, and energy efficiency. Through extensive computer simulations, we show that the proposed R-ACB technique outperforms the conventional ACB schemes.
All Science Journal Classification (ASJC) codes
- Signal Processing
- Information Systems
- Hardware and Architecture
- Computer Science Applications
- Computer Networks and Communications