Hybrid Tunable Flush Lag for High Availability and Performance
Original Publication Date: 2003-May-07
Included in the Prior Art Database: 2003-May-07
An algorithm is disclosed which accomplishes the following for a database system (e.g., a relational database): 1. Allow the user or system to restrict the amount of time needed for the database to become available again after a crash without significantly degrading performance. 2. Allow easy tuning, via a single parameter, of the system along a tradeoff continuum of maximum data availability (fastest restart after a crash) versus maximum performance (greatest transactional throughput). 3. Greatly reduce or eliminate throughput variance caused by checkpointing activity, such that checkpoint processing has little or no impact on throughput.