Evolution and Structure of the Internet
"Evolution and Structure of the Internet: A Statistical Physics Approach", by Romualdo Pastor-Satorras and Alessandro Vespignani. Cambridge University Press, 2004.
In my recent study on the structure of Internet, I happened to read most of this book. It is a very concise book on the structure of the Internet topology, its characterizing parameters, their statistical models, descriptions of topology creation tools available in the field and analysis of their performances. It is an easy read for the people who want to get the idea of the structure of the Internet as well as an impressive read for the people who want to understand it in a statistical point-of-view.
Internet can be visualized as a connected graph with the routers as the graphs node and their connections as the edges in the graphs. I want to short-list the defining characteristics of the Internet topology:
In my recent study on the structure of Internet, I happened to read most of this book. It is a very concise book on the structure of the Internet topology, its characterizing parameters, their statistical models, descriptions of topology creation tools available in the field and analysis of their performances. It is an easy read for the people who want to get the idea of the structure of the Internet as well as an impressive read for the people who want to understand it in a statistical point-of-view.
Internet can be visualized as a connected graph with the routers as the graphs node and their connections as the edges in the graphs. I want to short-list the defining characteristics of the Internet topology:
- degree: number of edges branches from a node in the Internet topology (property of nodes). Internet shows heavy-tail distribution in degree distribution which means that, while a huge number of nodes has very small degree, there are a small number of nodes with very large degree. Which also implies the "popular-get-more-popular" concept.
- shortest path length: the number of edges in the shortest paths between two end-points (property of node-pairs). Internet shows small world phenomena, that is, the average shortest path length in Internet is small.
- betweenness: the number of shortest paths goes through a particular node (property of nodes).
- clustering coefficient: an index indicating how well a node's neighbors are connected among themselves (property of nodes).
- scale-free behavior: it is a property of the topology as a whole. It says that, at whatever granularity we look at the Internet topology (AS-topology, router topology, intra-AS router topology), the properties of the topology follows the same statistical model.
- rich-club phenomena: this says that larger degree nodes are connected to each other with high probability.
- router-density is correlated with the population density. It is further correlated with the wealth of the regions.
- Logical Internet distance (hop count) does not correlate