Wide-Scale Internet Disconnection: Impact and Recovery on Social-Based P2P Overlays

In the recent years, there have been situations where wide-scale network fraction of Internet users have been disconnected from the rest of the network, due to natural hazards or national censorships (e.g., a government cuts connectivity to the outside world as a mechanism for suppression of uprisings). Peer-to-peer systems are known to be resilient in the presence of churn and uncorrelated failures. However, their behavior in extreme scenarios, where massive correlated failures occur, is not well-studied. In this paper, we consider nodes within a relatively small geographical region of the network (2.5 percent or fewer Internet users), and study whether users can communicate with their (n-hop away) social neighbors in a peer-to-peer fashion after network partitions. The contribution of our work is twofold: 1) we study the effect of such partitions on topology and routing of structured and social-based unstructured P2P overlays, including a newly proposed Social-aware overlay; 2) for the overlay with better routability observed from the first step, we propose and evaluate a social-based bootstrapping method for nodes to join the overlay, under the constraints imposed by firewalls and NATs. In our analysis, we consider both real and synthetic datasets of online social networks. The results of our evaluation show that structured P2P overlay routability is severely hampered by wide-scale partition events. In addition, the proposed social-based unstructured overlay network provides improved routability while maintaining a smaller number of links. Furthermore, in the evaluation of bootstrapping, by varying several input parameters, from 90 to 97 percent of nodes in Social-aware overlay could bootstrap and reach their direct-neighbors.