Document Type

Theses, Ph.D

Rights

This item is available under a Creative Commons License for non-commercial use only

Disciplines

Electrical and electronic engineering

Publication Details

Successfully submitted for the award of Doctor of Philosophy (Ph.D) to the Dublin Institute of Technology, 2014.

Abstract

In WLANs, the capacity of a node is not fixed and can vary dramatically due to the shared nature of the medium under the IEEE 802.11 MAC mechanism. There are two main methods of capacity estimation in WLANs: Active methods based upon probing packets that consume the bandwidth of the channel and do not scale well. Passive methods based upon analyzing the transmitted packets that avoid the overhead of transmitting probe packets and perform with greater accuracy. Furthermore, passive methods can be implemented locally or remotely. Local passive methods require an additional dissemination mechanism in order to communicate the capacity information to other network nodes which adds complexity and can be unreliable under adverse network conditions. On the other hand, remote passive methods do not require a dissemination mechanism and so can be simpler to implement and also do not suffer from communication reliability issues. Many applications (e.g. ANDSF etc) can benefit from utilizing this capacity information. Therefore, in this thesis we propose a new remote passive Capacity Utilization estimator performed by neighbour nodes. However, there will be an error associated with the measurements owing to the differences in the wireless medium as observed by the different nodes’ location. The main undertaking of this thesis is to address this issue. An error model is developed to analyse the main sources of error and to determine their impact on the accuracy of the estimator. Arising from this model, a number of modifications are implemented to improve the accuracy of the estimator. The network simulator ns2 is used to investigate the performance of the estimator and the results from a range of different test scenarios indicate its feasibility and accuracy as a passive remote method. Finally, the estimator is deployed in a node saturation detection scheme where it is shown to outperform two other similar schemes based upon queue observation and probing with ping packets.

DOI

10.21427/D7M02J

Share

COinS