This paper seeks to determine which attack strategies (hub, bridge, or fragmentation) are most effective at disrupting two online child pornography networks in terms of outcome measures that include density, clustering, compactness, and average path length. For this purpose, two networks were extracted using a web-crawler that recursively follows child exploitation sites. It was found that different attack strategies were warranted depending on the outcome measure and the network structure. Overall, hub attacks were most effective at reducing network density and clustering, whereas fragmentation attacks were most effective at reducing the network's distance-based cohesion and average path length. In certain cases, bridge attacks were almost as effective as some of these measures.
Kila Joffres, Martin Bouchard, Richard Frank, and Bryce Westlake. "Strategies to Disrupt Online Child Pornography Networks" European Intelligence and Security Informatics Conference (2011): 163-170.