Multi-Sensor Based MK/HyperK/1/M Queuing Model for Heterogeneous Traffic

Abstract

Integration of wireless sensor networks (WSNs) with Internet of Things (IoT) has given rise to applications such as Smart Home Networks, Vehicular Ad hoc Sensor Networks and Health and Wellness Wireless Sensor Networks among others. These applications tend to use multi-sensor nodes that have multiple sensors mounted on a single radio board and may require handling heterogeneous traffic. Modern WSNs are therefore no longer homogeneous in nature and have different quality of service (QoS) requirements. Modeling the performance of heterogeneous data is a challenging task owing to the complexity involved. In this paper we propose a MK/HyperK/1/M queuing model for multi-sensor nodes. Queues with Poisson arrivals and hyper-exponentially distributed service time and their applications have been widely discussed in literature however hardly any work can be found that uses them for modeling WSNs. In this paper we consider that the traffic from multi-sensor node is emanating from different sensor types within the node and is hence heterogeneous. The model captures the impact of varying arrival rates, mean packet lengths, buffer size and duty cycle on throughput. The model can be used with any duty cycle-based medium access control (MAC) protocols for WSNs. To test the model we apply it to IEEE 802.15.4 in beacon-enabled mode with star topology. Analytical results are corroborated by ns-2.35 simulations. Results indicate that the heterogeneity of traffic has a significant impact on the throughput and this may be useful in improving the QoS of applications.