Pangaea is a wide-area file system that targets the needs for data access and sharing of distributed communities of users. It federates commodity computers provided by users. Pangaea is built on three design principles: 1) pervasive replication to provide low-access latency and high availability, 2) randomized graph-based replica management that adapts to changes in the system and conserves WAN bandwidth, and 3) optimistic consistency that allows users to access data at any time, from anywhere.
The evaluation of Pangaea shows that Pangaea is as fast and as efficient as other distributed file systems, even in a LAN. The benefits of pervasive replication and the adaptive graph-based protocols become clear in heterogeneous environments that are typical of the Internet and large intranets. In these environments, Pangaea outperforms existing systems in three aspects: access latency, efficient usage of WAN bandwidth, and file availability.