Abstract: A data-replication system replicates a source database to a target database by first intelligently identifying and prioritizing a set of hot queries that have each in the past been submitted to the source database at a rate exceeding a threshold submission rate. Any previously replicated content contained in a database table accessed by a hot query is copied to a set of memory-resident blocks. The blocks are then updated with incremental changes made to the table's contents since the most-recent previous replication. Each updated block is copied to the target database in order of the priority of the block's corresponding hot query. When all blocks that contain data of a particular table have been copied into the target database, that table becomes available to users of the target database despite the fact that the entire database has not yet been fully replicated.

Abstract: A data-replication system replicates a source database to a target database by first intelligently identifying and prioritizing a set of hot queries that have each in the past been submitted to the source database at a rate exceeding a threshold submission rate. Any previously replicated content contained in a database table accessed by a hot query is copied to a set of memory-resident blocks. The blocks are then updated with incremental changes made to the table's contents since the most-recent previous replication. Each updated block is copied to the target database in order of the priority of the block's corresponding hot query. When all blocks that contain data of a particular table have been copied into the target database, that table becomes available to users of the target database despite the fact that the entire database has not yet been fully replicated.