Data Synchronization: A Complete Theoretical Solution for Filesystems

Data reconciliation in general, and filesystem synchronization particular, lacks rigorous theoretical foundation. This paper presents, for the first time, a complete analysis of two replicas filesystem. Synchronization has main stages: identifying conflicts, resolving them. All existing (both practical) synchronizers are operation-based: they define, using some rationale or heuristics, how conflicts to be resolved without considering effect resolution on subsequent conflicts. Instead, our approach is declaration-based: we define what constitutes all each possible scenario prove existence sequences operations/commands which convert into common synchronized state. These consist operations rolling back local changes, followed by performed other replica. The set rolled-back provides user with clear intuitive information proposed so she can easily decide whether accept them ask alternatives. states described specifying partial order describing need resolved, as well decision Using this classification, outcomes different conflict policies investigated easily.