Compact network alignment with mitigated sensitive information exposing in P2P networks: a community partition-based approach

Rui Tang, Yiming Peng, Jingxi Li, Xingshu Chen, Yi Zhang, Xian Mo

Abstract: The interaction data in P2P networks can be represented as graphs. When numerous nodes interact across various P2P applications, network alignment could reveal the potential for exposing users’ privacy or sensitive device information. Traditional structure-based network alignment techniques, which rely heavily on matrix operations of global adjacency matrices, are effective but not scalable for large networks due to high computational and memory demands. To address this challenge, we propose a community partition-based network alignment method (CPNA) that reduces space complexity by leveraging community information. This approach transforms the complex computations required for global network alignment into smaller, more feasible matrix operations within these communities. The effectiveness of the CPNA method is validated through experiments on two distinct datasets, demonstrating that it maintains sufficient precision while decreasing the space complexity for network alignment.