Previous study of real-time databases mainly focuses on scheduling and concurrency control of single type real-time transactions. However, there is a very common demand on processing mixed transactions. This paper first introduces a two-level real-time database system model, which provides the support of using aperiodic task scheduling algorithms to improve the system performance. Next, a new real-time concurrency control protocol called mixed concurrency control with dynamic adjustment of serialization order using timestamp intervals (MCC-DATI) is proposed to ensure the data consistency among mixed transactions. The schedulability of hard real-time transactions can be guaranteed by adopting dynamic priority driven scheduling algorithm and bounding the blocking time from aperiodic soft real-time transactions, and the deadline miss ratio of soft real-time transactions may be reduced by adopting aperiodic task scheduling algorithm and dynamic adjustment of serialization order using timestamp intervals. Simulation experiments show that the MCC-DATI protocol can improve the system performance under different workloads and deadline constraints, as compared with previous concurrency control protocols of mixed transactions.