Abstract
Data integrity is one aspect of database consistency concerned with the validity of databases. Integrity requirements are expressed by means of a set of integrity constraints. A database state is consistent (with respect to a set of constraints) if all these constraints are evaluated to be true in it. An execution preserves integrity if it transfers a database from an consistent state to another consistent state and each transaction sees an consistent state in the execution.
Database integrity is usually enforced through integrity enforcement mechanisms [BM88] or correct scheduling of transactions. The latter approach, however, works only for those constraints that one can determine statically (i.e., independently of database state) whether they are preserved by a single transaction. In this paper, we are interested in the second approach and assume that each transaction, when executed alone, preserves all constraints.
Database integrity is usually enforced through integrity enforcement mechanisms [BM88] or correct scheduling of transactions. The latter approach, however, works only for those constraints that one can determine statically (i.e., independently of database state) whether they are preserved by a single transaction. In this paper, we are interested in the second approach and assume that each transaction, when executed alone, preserves all constraints.
| Original language | English |
|---|---|
| Publication status | Published - 1990 |
| Externally published | Yes |