D-SeDGAM: A Dynamic Service Differentiation Based GTS Allocation Mechanism for IEEE 802.15.4 WSN

Abstract

The plethora of applications for Wireless Sensor Networks (WSN) has experienced significant growth in recent years. Every application has a different set of requirements and it is highly desirable for all of them to share the available resources in an optimized way. The IEEE 802.15.4 standard for WSN includes guaranteed time slots (GTS) to schedule QoS demanding applications. However, the standard does not provide any type of QoS GTS scheduling mechanism and the authors are unaware of any other work that provides for service differentiation based GTS scheduling. We propose DSeDGAM, a novel dynamic service based differentiation GTS allocation mechanism to provide QoS and service based differentiation for WSN applications. Our solution takes into account static and transient aspects of QoS provisioning in WSN to construct a dynamic scheduling method. This work demonstrates that D-SeDGAM provides significantly better performance in terms of QoS than the standard allocation method by simultaneously satisfying QoS and service differentiation demands.

Enhancement of the Traffic Differentiation Architecture for WBAN Based on IEEE 802.15.4

“…To meet this need, other protocols such as D-SeDGAM [35] allow the differentiation between services. This class suffers from several limits.…”

Traffic differentiation and dynamic duty cycle adaptation in IEEE 802.15.4 beacon enabled WSN for real-time applications

“…A dynamic service differentiation scheme was proposed in [9] called D-SeDGAM, based on the static and transient aspects of QoS provisioning in WSN. The dynamic priority algorithm takes into account the delay, buffer and bandwidth metrics to prioritise nodes and construct the GTS scheduler.…”

“…Nevertheless, three problems have not been addressed in employing IEEE 802.15.4 beacon enabled mode for WSN applications. First, all the work outlined above except for [9] segregate the evaluation of CAP and CFP for real-time applications rather than resolving it as a complete superframe utililising both CAP and CFP. Secondly, adapting the duty cycle using the superframe structure (SO and BO) is desirable to accommodate the changing context of the WSN traffic; however, it is always assumed fixed in existing approaches after the network starts.…”

“…In voice communication, interactive voice requires an endto-end latency of 250 ms or less, beyond which users notice a drop in interaction quality [15]. Other literature recommends that latency in WSN for real-time traffic should not exceed 100 ms [9,16]. For this case we use 250 ms for the emergency delay threshold.…”

“…4.3. Note that exponential distribution is typically adopted by most simulation and analytical studies on CSMA/CA [9,18,19].…”