Performance analysis of resource sharing in wireless networks: Analytical and empirical perspectives.

摘要

In wireless networks, the efficient sharing of scarce wireless spectral resources is important in order to provide guaranteed Quality-of-Service (QoS) to the wireless users. The effectiveness of resource sharing schemes in wireless networks are often heavily influenced by different aspects of the system behavior, such as user mobility, traffic dynamics and practical realization constraints. In this thesis, using analytical modeling and empirical measurement techniques, we investigate the impact of these system behaviors on the performance of resource sharing in wireless networks. In particular, we investigate the dynamic sharing of an access point's bandwidth resources among moving vehicles in a vehicular network, the adaptive sharing of the medium access resources among nodes with different and varying traffic loads in a wireless sensor network, and the practical implementation of network resources sharing among users and applications with different QoS requirements in 3G wireless networks.In the first part of this thesis, we focus on Drive-thru Internet systems where access points (AP) are placed on roadsides and vehicles passing through the coverage range of the APs can download data from them. The amount of data downloaded by an individual user is affected not only by the scheduling algorithms, but also by the user dynamics, i.e. the movement of the vehicles which impacts the amount of time the vehicle spends in the AP's coverage range, as well as the number of contending vehicles for the AP's resources. We have developed practical analytical models with tractable solutions to characterize the data communication performance of a vehicle in a Drive-thru Internet system. A distinctive aspect of our models is that they combined both vehicular traffic theory and wireless network/protocol properties to investigate the effects of various system parameters on a drive-thru vehicle's data communication performance.In the second part of this thesis, we examine resource sharing in wireless sensor networks in terms of the node access to the wireless medium. We propose an energy-efficient TDMA-based MAC protocol that significantly reduces energy consumption in the network, while efficiently handling network traffic load variations and optimizing channel utilization through a timeslots stealing mechanism and timeslots reassignment procedure. We have analytically derived the average delay performance of our MAC protocol, with and without the timeslots stealing feature. Our delay model, validated via simulations, shows that the timeslots stealing feature can substantially improve the protocol throughput in situations with varying and asymmetric traffic patterns. Simulation results show that the timeslots reassignment procedure is efficient in handling the longer timescale changes in the traffic load, while the timeslots stealing mechanism is better in handling the shorter timescale changes in the traffic patterns.The third part of this thesis focuses on our empirical investigations into the performance of practical implementation of resource sharing schemes in 3G wireless networks. We have investigated the performance of multiple commercial 3G networks in Hong Kong, in terms of their ability to provide service guarantees to different traffic classes as well as the fairness of the radio-link scheduler in allocating the bandwidth resources to multiple data calls in a saturated network. We have also investigated the data throughput, latency, video and voice calls handling capacities of the 3G networks under saturated network conditions. Our findings point to the diverse nature of the network resources allocation mechanisms and the call admission control policies adopted by different operators. Our results also show that the 3G network operators seem to have extensively customized their network configurations in a cell-by-cell manner according to the individual site's local demographics, projected traffic demand and the target coverage area of the cell. As such, the cell capacity varies widely not only across different operators but also across different measurement sites of the same operator.