Temporal Interoperability in Multi+Temporal Databases

Abstract

A multidatabase system allows applications and users to access heterogeneous information through a common interface. Information is shared and exchanged in a transparent way within a federation of autonomous database systems. In a multi+temporal database, the connected systems may be standard (snapshot) databases, temporal databases (either valid-time, transaction-time or bitemporal) or even single multitemporal systems in which data with different temporal formats may coexist. In this paper we consider semantic interoperability problems arising from the interaction of temporally heterogeneous data in a multi+temporal environment. In particular, we focus on the evaluation of multitemporal algebraic expressions for the support of an SQL-like query interface. The basic requirement of a multitemporal algebra is that all the operations involved must be carefully designed in order to retrieve the largest amount of certain stored information and avoid the creation of spurious information during the interaction of heterogeneous data. The main difficulties underlying the definition of such operations are presented in the paper, together with solutions to overcome them. They include conversion of data from one temporal format into another, definition of lossless operations on temporally incomplete information, and semantic optimization of multitemporal algebraic expressions.