Multiversion Serializability and Concurrency Control
Multiversion concurrency control; Multiversion concurrency control algorithms; Multiversion databases
Given a multiversion database, where each data item is a sequence of its versions, the number of versions of a data item may be limited or not. If it is unlimited, then each update of a data item over the limit gives rise to its next version. If it is limited, than each update of a data item replaces its oldest version. In case of limited number of versions, a database is called a K-version database. In multiversion databases any read operation of a data item, subsequent to a write operation of this data item, may access any of its currently existing versions. Thus, a multiversion schedule of a transaction set differs from the ordinary, mono-version schedule by a mapping of the data item read operations into the data item version read operations. Multiversion serializability plays the same role for the multiversion databases, as serializability for the ordinary,...
- 1.Bernstein PA, Goodman N. A sophisticate’s introduction to distributed database concurrency control. In: Proceedings of the 8th International Conference on Very Data Bases; 1982. p. 62–76.Google Scholar
- 2.Bernstein PA, Goodman N. Concurrency control and recovery for replicated distributed databases. Technical Report TR-20/83, Harvard University; 1983.Google Scholar
- 4.Bernstein PA, Hadzilacos V, Goodman N. Concurrency control and recovery in database systems. Reading: Addison-Wesley; 1987.Google Scholar