Abstract: The input of a computer executable process, is logically subdivided, without reading, into a plurality of partitions which are distributed to a plurality of processors in which respective subtasks including the reading of those partitions, are carried out. The method allows distribution of processing of a large amount of data to a plurality of processors cooperating in a way that the load imposed on each processor is proportional to its capacity to do the work.
Abstract: The log file maintained by a DBMS is used, possibly in conjunction with hardware that listens to the communication between a computer and the storage controller to create cache buffers and a locking mechanism that enable applications running on one computer system to consistently access the data maintained and updated by a different computer.
Abstract: The input of a computer executable process, is logically subdivided, without reading, into a plurality of partitions which are distributed to a plurality of processors in which respective subtasks including the reading of those partitions, are carried out. The method allows distribution of processing of a large amount of data to a plurality of processors cooperating in a way that the load imposed on each processor is proportional to its capacity to do the work.
Abstract: A change propagation method for a migration from a source table to a target table in a DBMS which uses the log file available in the DBMS for recovery from failures. From this log file a last data-independent key is derived to represent an immediate previous migration and a condense table is generated with records representing before and after images of a data row which has changed in the source table. For each source table in a migration definition files containing delete statements, update statements, an insert table and an ignore file are provided. A target table is updated based upon commands in an insert statement file generated from the migration statement, the ignore file, the source table and the insert table.