Enhancing real-time DBMS performance with multiversion data and priority based disk scheduling

Real-time multiversion concurrency control algorithms are proposed, to: increase concurrency, adjust the serialization order dynamically and work without an estimate of a transaction´s runtime. The authors also propose disk scheduling algorithms which consider not only the transactions which request input/output (I/O) but also those affected by I/O. They consider transactions which are directly affected when priorities of I/O requests are assigned, in addition to transactions which generates these requests. The real-time disk resident database system model and the multiversion concurrency control algorithms are described. The real-time disk scheduling algorithms and their properties are also described. Real-time multiversion concurrency control and disk scheduling algorithms are shown to decrease miss ratio significantly