Survey of data replication in P2P systems

摘要: Large-scale distributed collaborative applications are getting common as a result of rapid progress in technologies (grid, peer-to-peer, and mobile computing). Peer-to-peer (P2P) systems particularly interesting for they can scale without the need powerful servers. In P2P systems, data storage processing across autonomous peers, which join leave network at any time. To provide high availability spite such dynamic behavior, rely on replication. Some replication approaches assume static, read-only (e.g. music files). Other solutions deal with updates, but simplify replica management by assuming no update conflicts or single-master (i.e. only one copy replicated accepts write operations). advanced applications, must semantically rich XML documents, relational tables, etc.) using high-level SQL-like query language, likely to more sophisticated capabilities multi-master all replicas accept operations) conflict resolution. These issues addressed optimistic Optimistic allows asynchronous updating so that even though some nodes disconnected have failed. As result, users collaborate asynchronously. However, concurrent updates may cause divergence conflicts, should be reconciled. this survey, we present an overview replication, focusing approach provides good properties environments. We also introduce implement. particular, show current do not eventual consistency among presence apart from APPA system, system building.