Abstract: A bidirectional database replication system includes a plurality of nodes. Each transaction at an originating node is paused prior to a commit operation. Ready to commit tokens are sent to the other nodes in the system to determine if the other nodes are prepared for the commit operation for the paused transaction. If all of the ready to commit tokens properly return to the originating node from the other nodes, thereby indicating that the other nodes are prepared for the commit operation, then the transaction is commited. For lengthy transactions, ready to sync tokens are assigned at one or more predesignated intermediate points in the transaction, and propagate throughout the system in a similar manner. The transaction continues to execute as long as all ready to sync tokens properly return to the originating node. The pause-before-commit and sync point schemes are used to avoid collisions at any of the nodes.

Abstract: Processors are assigned in a multiprocessor environment to a plurality of processes that are executed in the multiprocessor environment. Each process has a process pair defined by a primary process that executes on a first processor and a backup process that executes on a second processor. The processors are in communication with one another via a communication network. The plurality of processors are organized into predefined processor pairs. Every process pair is assigned to one of the processor pairs, wherein no process pairs span across different processor pairs.

Abstract: A method is provided for splitting a system comprising a processor subsystem including a plural number (np) of processors, and an instance of a database. A number (nn) of nodes is selected that the system is to be split into. Each node may be provided with a processor subsystem including at least one processor. At least some of the nodes further include an instance of at least a portion of the database. A network is provided for communications among the nodes. At least some of the nodes that contain an instance of at least a portion of the database are automatically synchronized so that contents of the database that exist in plural nodes remain synchronized in at least some of the plural nodes whenever contents of the database at one of the nodes is changed. Systems with multiple processors, but without databases, may also be split in the same manner.

Abstract: In a database synchronization process, a block of data is read from a source database at the first node. A marker is assigned to the block of data, and the marker is then written to an audit trail at the first node. The marker is sent from the audit trail to a second node having a target database. The block of data is sent from the first node to the second node without passing the block of data through the audit trail. At the second node, the block of data is stored in the target database upon receipt at the second node of the block of data and the marker assigned to the block of data. This process is repeated for additional blocks of data in the source database, wherein a marker is assigned to each subsequent block of data. In this manner, some or all of the source database may be replicated to the target database without having to store or pass any data in the source database in or through the audit trail at the first node.

Abstract: In a database synchronization process, a block of data is read from a source database at the first node. A marker is assigned to the block of data, and the marker is then written to an audit trail at the first node. The marker is sent from the audit trail to a second node having a target database. The block of data is sent from the first node to the second node without passing the block of data through the audit trail. At the second node, the block of data is stored in the target database upon receipt at the second node of the block of data and the marker assigned to the block of data. This process is repeated for additional blocks of data in the source database, wherein a marker is assigned to each subsequent block of data. In this manner, some or all of the source database may be replicated to the target database without having to store or pass any data in the source database in or through the audit trail at the first node.

Abstract: Database replication systems replicate blocks of transaction steps or operations with synchronous replication, and perform dual writes with queuing and blocking of transactions. Tokens are used to prepare a target database for replication from a source database and to confirm the preparation. Database replication systems switch between a synchronous replication mode and an asynchronous replication mode, and then back to a synchronous replication mode, based on detection of selected events.

Abstract: Transaction ping-pong is selectively prevented in a bidirectional database replication system. The system has a plurality of nodes connected via communication media in a topology. Each node includes a database and a transaction transmitter or collector which sends transactions posted to the database to a database at one or more other nodes for replication in the databases of the one or more other nodes. All transactions to be posted to databases in remote nodes that were sent by a local node are detected, and the database at the local node is inhibited from posting selective transactions which were detected as being originally sent by the local node.