Contribution to the theory of concurrency in databases
A database can be seen as a set of entities which satisfy certain assertions called integrity constraints. A database is simultaneously accessed by numerous transactions. Each transaction is composed of a sequence of atomic actions and performs a set of operations. In order to guarantee the good execution of all the operations and the integrity of the database, it is necessary to limit the simultaneous execution of transactions.
In current database management systems locking logical or physical parts of the database during the execution of each transaction is performed. This solution restricts the degree of sharing. In this paper, a better characterization of possible schedules is proposed. Then, scheduling algorithms avoiding incorrect executions are proposed.
Unable to display preview. Download preview PDF.
- 3.R. W. ENGLES "Currency and Concurrency in the Cobol Database Facility" IFIP Working Conference on Modeling in DBMS North-Holland Pub. Co., Amsterdam 1976, pp 339–363Google Scholar
- 4.S. TODD "Automatic Constraint Maintenance and Updating Defined Relations" IFIP Congress 77 — Toronto — Août 1977 — pp 145–148Google Scholar
- 5.G. GARDARIN, P. LEBEUX "Scheduling algorithms for avoiding inconsistency in large databases" Third International Conference on Very Large Databases Proc. edited by IEEE, New-York 1977 — pp 501–506Google Scholar
- 6.J. B. ROTHNIE, N. GOODMAN, P. A. BERNSTEIN "The Redundant Update Methodology of SDD.1: A System for Distributed Databases" Technical Report CCA-77-02 June 15, 1977, 70 pGoogle Scholar
- 7.G. GARDARIN, S. SPACCAPIETRA "Integrity of Databases: a General Lockout Algorithm with Deadlock Avoidance" IFIP Working Conference on Modeling in DBMS North-Holland Pub. Co., Amsterdam 1976 pp 395–411Google Scholar
- 8.CODASYL Programming Language Comittee "Data Base Task Group Report", April 1971 "Data Base Facility Proposal", March 1973Google Scholar