Abstract: An Internet-scale file sharing system includes a client-side file sharing application that allows file-sharing users to identify files to share and transmit metadata corresponding to those files to a metadata repository. A server-side application operating on the metadata repository tracks metadata received from associated file-sharing users, as well as metadata from other affiliated metadata repositories. Each metadata repository acts as a search engine for any querying users and can provide search results based on locally stored metadata alone. Each metadata repository may additionally choose to locally-store popular files from an associated file-sharing user so as to alleviate transmission burdens on that file-sharing user. Associated metadata repositories each periodically synchronize their stored metadata so that all metadata repositories may generate similar search results.