Topological characteristics of IP networks

Topological analysis of the Internet is needed for developments on network planning, optimal routing algorithms, failure detection measures, and understanding business models. Accurate measurement, inference and modelling techniques are fundamental to Internet topology research. A requirement towards achieving such goals is the measurements of network topologies at different levels of granularity. In this work, I start by studying techniques for inferring, modelling, and generating Internet topologies at both the router and administrative levels. I also compare the mathematical models that are used to characterise various topologies and the generation tools based on them. Many topological models have been proposed to generate Internet Autonomous System(AS) topologies. I use an extensive set of measures and innovative methodologies to compare AS topology generation models with several observed AS topologies. This analysis shows that the existing AS topology generation models fail to capture important characteristics, such as the complexity of the local interconnection structure between ASes. Furthermore, I use routing data from multiple vantage points to show that using additional measurement points significantly affect our observations about local structural properties, such as clustering and node centrality. Degree-based properties, however, are not notably affected by additional measurements locations. The shortcomings of AS topology generation models stems from an underestimation of the complexity of the connectivity in the Internet and biases of measurement techniques. An increasing number of synthetic topology generators are available, each claiming to produce representative Internet topologies. Every generator has its own parameters, allowing the user to generate topologies with different characteristics. However, there exist no clear guidelines on tuning the value of these parameters in order to obtain a topology with specific characteristics. I propose a method which allows optimal parameters of a model to be estimated for a given target topology. The optimisation is performed using the weighted spectral distribution metric, which simultaneously takes into account many the properties of a graph. In order to understand the dynamics of the Internet, I study the evolution of the AS topology over a period of seven years. To understand the structural changes in the topology, I use the weighted spectral distribution as this metric reveals differences in the hierarchical structure of two graphs. The results indicate that the Internet is changing from a strongly customer-provider oriented, disassortative network, to a soft-hierarchical, peering-oriented, assortative network. This change is indicative of evolving business relationships amongst organisations.

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

UCL (University College London)

Year dc:date.issued

2008

Author dc:creator

• Haddadi, H.

• Department of Electronic and Electrical Engineering

Language dc:language

en, eng