Performance analysis of temporal queries

Temporal databases maintaining history data on line extend conventional databases with capabilities for historical queries and rollback operations. To analyze the performance of temporal queries on databases using various access methods, we propose a model that takes a temporal query and a database schema as input, and outputs the estimated I/O cost for the query on that database. The model consists of four transformations through a series of formal expressions characterizing all phases of query processing. We validate the model by comparing the I/O cost estimated from the model with the actual cost measured from a prototype temporal DBMS. Since conventional databases are a subset of temporal databases, the model can also be used to analyze the performance of conventional databases.