Integrating checkpointing with transaction processing

Many industrial applications consist of different components that were designed and implemented separately. Software reuse and component software architecture often dictate this style of building applications out of existing modules. In such an application, some modules may use databases to store critical data, while other components may use regular files or volatile memory either by design or to improve performance. Therefore, global coordination is necessary for maintaining data consistency and recoverability across multiple datastores. We describe a way of integrating check-pointing with transactions to ensure consistency and recoverability of data stored outside the databases. Our approach transforms an existing checkpoint library into a transactional resource manager that can participate in such a global coordination.