Designing File Replication Schemes for Peer-to-Peer File Sharing Systems

Peer-to-peer (P2P) file sharing systems are becoming increasingly popular due to their flexibility and scalability. We propose a new model to design file replication schemes for P2P file sharing systems. The model introduces expected costs for serving user requests for the files which are computed from node up/down statistics. Based on the model we introduce and develop several methods to determine the sets of nodes to store copies of the files in order to optimize certain performance metrics (e.g., maximize the system hit rate, minimize the total expected cost). We verify the effectiveness of the file replication schemes via simulation. We also outline a framework to implement the file replication schemes for P2P file sharing systems in a distributed and adaptive manner. The framework scales to a large number of nodes and files and can handle user request pattern change via file migration.