In a real-time database system, timing constraints are associated with transactions and data accessed by transactions may be valid for only specific time intervals. A correct transaction processing means that the transaction completes tis processing before its timing constraints and uses data that are both absolutely and relatively timing-consistent. We define the two distinct performance indeces as timing constraint and data consistent constraint. It is ideal that an algorithm both improve the two indeces at the same time , but they may affect each other. In this paper, we explore that whether we can improve them simultaneously and, if not, the trade-off. The simulation results show that by taking advantage of the data temporal consistency constraints in transaction scheduling improve significantly the system performance (miss percentage and temporal inconsistency percentage). The results also show that some system parameters, such as the period of a transaciton and the relative validety interval of a data set, have strong effects on the sysem performance.
展开▼
