Studies in applications of queueing models to wireless sensor networks

Abstract

Wireless sensor networks have tremendous potential for numerous application areas, including environmental, medical, military, transportation, entertainment, crisis management, homeland defenses and smart spaces. The design of a wireless sensor network depends significantly on the application and must consider factors such as environment, the application s design objectives, cost, hardware, power, limited storage, energy efficiency etc. Energy saving factor is a major concern in wireless sensor networks since the sensors are battery operated and hence it is important to reduce or completely remove the battery replacement operation. Further, as sensing of the signals and communicating to other nodes consume more power, it is necessary to minimize the energy consumption in such networks. Mathematical analysis of Queueing models in wireless communication network helps to study factors such as energy saving schemes, channel control, quality of service etc. The focus of this thesis is on queueing models applied to wireless sensor networks with special emphasis on energy saving factor. In this work, various queueing models incorporating different features have been studied. In all the models, the transient and steady state probabilities and steady state performance measures with respect to maintenance mode, sleep mode and active mode subject to catastrophe are obtained; when a catastrophe strikes the system, it is assumed that all the customers are wiped out from the system and the system enters the maintenance mode. Further, customers are not permitted to join the system when the system is in maintenance mode. To reduce power consumption of the system, the concept of sleep mode and an idle state as part of active mode is introduced. An M/M/1 queue with the server operating in three modes - maintenance mode, sleep mode and active mode which comprises idle state and busy state subject to catastrophes has been studied. At the end of the maintenance mode, the server returns to the sleep mode with probability 1- p and ent