The study of network attacks by using complex network models

Abstract

A complex network is a network with non-trivial topological features, which means the pattern of connections between pairs of nodes in the network is neither purely regular nor purely random. It describes a wide range of networks in nature and society, including biological networks, social acquaintance networks and the human-made transport and communication networks like the Internet. In recent years, modelling and analysis for different types of complex networks has become an interdisciplinary subject and many characteristics of complex networks have been discovered and proved, such as they are sensitive to targeted attacks which may lead to fracturing the whole network. That means to reduce or even avoid this kind of attacks whenever possible is important for maintaining the networks. One approach to achieve it is to optimize the routing algorithm aimed at reducing the traffic of those backbone-nodes. Doing a literature review is the first step of the project in order to get a deep understanding about complex networks and figure an outline of predecessors’ studies. On the basic of literature review, computer simulations in JAVA language are carried out to generate and analyse a scale-free network in BA model — a typical complex network in one of the most famous complex network models. After that, network transport processes with shortest path first (SPF) routing algorithm and an optimized algorithm are simulated. According to comparisons between these two algorithms in term of transport performance, a new routing algorithm — parallel routing algorithm, which is the combination of previous two algorithms and shows quite satisfying transport performance, is raised and discussed.