In a community cloud, multiple user groups dynamically share a massive number of data blocks. The authors present a new associative data sharing method that uses virtual disks in the MeePo cloud, a research storage cloud built at Tsinghua University. Innovations in the MeePo cloud design include big data metering, associative data sharing, data block prefetching, privileged access control (PAC), and privacy preservation. These features are improved or extended from competing features implemented in DropBox, CloudViews, and MySpace. The reported results support the effectiveness of the MeePo cloud.