More and more data of large and integrated business applications are stored in (relational) database format. Nowadays, a database size of one hundred gigabytes or more is not uncommon anymore. Moreover, due to legal reasons or business requirements of an enterprise data must very often be kept ("archived") over a very long period of time, such as 10 or 20 years or even more. A database archive may therefore easily grow to terabytes within just a couple of years. Of course, the data must not only be archived for years but must also still be accessible in reasonable time for auditing purposes, reuse, trend analysis, etc. In this paper we present a proposal for data archiving functionality which is fully integrated in the DBMS "under the cover" of an extended SQL database language interface. We give a motivation for such an integrated database archive, also in contrast to other, nonintegrated approaches, and we present SQL extensions for the DDL and DML part to handle both current and archived data in a seamless fashion. A discussion of implementation aspects leads to special problems of integrating tertiary storage into a DBMS which is considered a must for the management of large amounts of data.
@TECHREPORT{Her95a,
AUTHOR = {A. Herbst and K. K{\"u}spert and R. Schaarschmidt},
TITLE = {Datenarchivierungsfunktionalit{\"a}t in der
{SQL{\it x}}-{N}orm},
TYPE = {Forschungsergebnisse der Fakult{\"a}t f{\"u}r Mathematik
und Informatik},
NUMBER = {Math/95/8},
INSTITUTION = {Institut f{\"u}r Informatik},
ADDRESS = {Friedrich-Schiller-Universit{\"a}t Jena},
MONTH = jul,
YEAR = {1995}
}