A new framework for dynamically evolving database environments
| Published date | 01 May 1999 |
| DOI | https://doi.org/10.1108/EUM0000000007141 |
| Pages | 144-158 |
| Date | 01 May 1999 |
| Author | E.J. Yannakoudakis1,C.X. Tsionos,C.A. Kapetis |
| Subject Matter | Information & knowledge management,Library & information science |
A NEW FRAMEWORK FOR DYNAMICALLY EVOLVING
DATABASE ENVIRONMENTS
E.J. YANNAKOUDAKIS1, C.X. TSIONOS
and C.A. KAPETIS
eyan@aueb.gr
Department of Informatics, Athens University of Economics and Business
76, Patission Street, Athens 10434, Greece
This paper describes research work carried out with the aim to
investigate dynamically evolving database environments and corre-
sponding schemata, allowing storage and manipulation of variable
length data, a variable number of fields per record, variable length
records, manipulation of authority records and links between
records and fields, and dynamically defined objects (relations in the
traditional sense). The paper proposes a new framework for the
definition of a unified schema that eliminates completely the need
for reorganisation at both logical and internal levels. Retrieval of
data is optimised through self-contained storage chunks that also
vary dynamically.
1. INTRODUCTION
It is well known that database models and database management systems do not
handle changes to the underlying schemata (logical and internal) very well.
Although the schema of a database is usually considered to be stable, in practice
it changes due to two main reasons: (a) the real world which is modelled by the
database is continuously changing and (b) the database designers often make mis-
takes in identifying all the attributes and the relationships between them during
the construction of the database. Typical changes to the schema are the addi-
tion/deletion of a column, the modification of a 1:1 relationship to 1:M or even
M:M, the projection of a column, the creation or dropping of a table etc.
Although changes to the schema do not take place as frequently as changes to
the actual data, when they do take place they require special attention because the
slightest modification affects applications directly. Frequently, the source code of
programs is scanned in order to locate all the sections that need to be modified.
Sjøberg [1] has proposed a tool that helps this task, but the effort remains consid-
erable. For example, let’s suppose that a field is projected to a separate table and
a reference to it has to be kept in the original table. This rather simple schema
change requires that every select, insert, update and delete statement regarding
the original table be investigated thoroughly. This problem is clearly related to
schema evolution which remains a research topic [2].
There are mainly two approaches to schema evolution. The first uses the tem-
poral database theory and incorporates temporal support within the meta-database,
144
Journal of Documentation, Vol. 55, No. 2, March 1999
© Aslib, The Association for Information Management.
All rights reserved. Except as otherwise permitted under the Copyright, Designs and Patents Act
1988, no part of this publication may be reproduced, stored in a retrieval system, or transmitted in
any form or by any means, electronic, mechanical, photocopying or otherwise without the prior
written permission of the publisher.
Aslib, The Association for Information Management
Staple Hall, Stone House Court, London EC3A 7PB
Tel: +44 (0) 171 903 0000, Fax: +44 (0) 171 903 0011
Email: pubs@aslib.co.uk, WWW: http://www.aslib.co.uk/aslib
144
1Author for all correspondence
Journal of Documentation, vol. 55, no. 2, March 1999, pp. 144–158
To continue reading
Request your trial