Rich-Club Connectivity in Large-Scale Complex Networks

In the real-world network, i.e., Internet, the nodes with high degrees are found closely connected with each other, which is stated as the rich-club connectivity. Obtaining the rich club connectivity is an easy job for small graphs. However, with the continuous expansion of the network size, especially for the online social networks, the naive algorithm could not handle these complex networks any more for the reason of limited memory and too much consuming time. In order to tackle this problem, in this paper, we propose two models for computing rich-club connectivity on MapReduce. We evaluate two models and find the second model outperforms the first one on large scale networks. Then we employ the second model to obtain the rich-club connectivity of several large-scale networks, including both technological and social ones. It is also interesting that as compared to technology networks, the rich-club connectivity in large-scale social systems is indeed different.