The unprecedented growth of data at geographically distributed locations coupled with tremendous improvement in networking capabilities over the last decade strongly motivate the need for efficient data management in widearea network (WAN) environments such as Peer-to-Peer (P2P) networks and GRIDs. In particular, data availability and performance demands on WAN applications are now greater than ever before. While replication has been traditionally used for maximizing both data availability and performance, this paper contends that replication schemes for traditional distributed environments (e.g., clusters) do not adequately address the requirements of WAN environments. Notably, issues such as node heterogeneity (in terms of processing capacity and available disk space for storing replicas), significant variations in bandwidth, lack of centralized control, lack of global knowledge, distributive ownership and scalability make replication in WAN environments significantly more challe...