Impact of Quality of Information on Performance and Scalability of Wireless Networks PDF Download

Author: Scott Rager
Publisher:
ISBN:
Category :
Languages : en
Pages :

Get Book

Book Description
This dissertation centers around two relatively new fields of study in communication networks. The first, Quality of Information, or QoI, refers to the inherent and contextual characteristics of data that define its utility to a particular application. The second, wireless ad hoc networks, are a field of network applications in which wireless devices, which are often mobile, must exchange information without the support of existing infrastructure, making them highly suitable for applications like emergency networks, military tactical networks, and sensor networks. Both QoI and wireless ad hoc networks bring a distinct set of design challenges to light because of their novelty and unique attributes. Both, however, also provide new opportunities for rethinking network research. Specifically, we expand our approach to designing network protocols to include algorithms that control a number of layers in the network stack jointly using current network conditions, such as backpressure routing. Within this approach to network control, we also seek to understand how the QoI metrics of accuracy, precision, and freshness of the network state information impact performance. Additionally, we consider the QoI of application layer data. We do so first by designing algorithms to select data in wireless sensor networks with the goal of maximizing the coverage of the selected data. We then model the scalability and limitations of networks when designing them to provide a minimum level of QoI instead of traditional approaches that simply offer asymptotic analysis.In each contribution of this dissertation, we aim to understand the impact of QoI on network control and analysis by building flexible models able to be applied to a variety of potential applications. In our analysis, we have strived to consider the practicality of applications in real-life scenarios. We achieve this practicality by considering overhead of protocols, developing and analyzing approximation algorithms to computationally complex algorithms, and providing realistic simulation results generated by the Qualnet and ns-3 network simulation software packages.We first study the framework of backpressure routing, providing analysis of its performance under delayed or inaccurate network state information as well as a close look at how overhead impacts performance in distributing this network state information. Next, we use Lyapunov optimization techniques from backpressure routing theory in a QoI-aware application to data selection for maximum information coverage with cost constraints in sensor networks. Finally, we use QoI to create a framework that outlines practical scalability and maximum achievable QoI limits. Overall, the goal of this document is to explore how Quality of Information impacts design and understanding of control algorithms and performance limits in ad hoc networks designed to support a variety of applications.