This thesis describes a new group of algorithms for scheduling real-time transactions, or transactions with deadlines. The algorithms are evaluated via a detailed simulation study. Our results show that under a wide range of workloads, the real-time algorithms perform significantly better than a conventional transaction scheduling algorithm. This thesis also studies the problem of scheduling disk requests with deadlines. Three new algorithms are proposed and their performance is evaluated via simulation. One real-time algorithm, FD-SCAN, was found to perform better than all algorithms tested, both real-time and conventional, under a wide range of experimental conditions.