Guest Editor's Introduction
Future Directions in
Database Systems-
Architectures for
Information Engineering
Nick Roussopoulos, University of Maryland
H as the true "information age" arrived? Earlier but we fall short of the whole thing. Summation or the
this decade, many people were assuring us that gathering of information does not do it.
H its "dawn" had arrived and that it would com- I dare to predict that the dawn will be definitely gone
pletely change our lives. The world would be knowl- when we all recognize that the knowledge database is
edge-driven by accessing accurate information in a the most fundamental component of any system and
snap. The enormous repositories of information and that all other components are engineered in and around
knowledge needed to design and use tomorrow's com- the management and control of information. It will be
plicated systems would be found by "letting your gone when the policy makers understand that a lift-off
fingers do the database walk." of a shuttle is not the physical movement per se, but the
In anticipation of this, there has been a lot more data control of all the events preceding it, the controlled
gathering. Even more than future information systems decisions that drive it, and the effects of those decisions
would be able to digest. Is data gathering sufficient to on future states of it. The physical movement is the
bring us out of the still gray period of dawn? The answer result of a set of controlled actions and is as important,
is no. Data gathering is useful only for "post-mortem or unimportant if you like, as the rotating movement of
analysis." The complexity of today's systems requires the wheel of a car during your search for the right exit
engineering information models capable of rapidly on a highway in New York City. Only the realization of
recognizing and reusing abstractions acquired from the importance of information engineering and ap-
massive amounts of experimental data and from propriate funding will bring us into the information age.
lengthy analyses performed on that data. In response to the information age vision, database
How long is dawn going to last anyway? Five years of system requirements have changed. The management
hardware development got you the processing power of of complex and continuously evolving factual knowl-
a computer center in a PC for less than $1000. After edge is very different from that of the "snapshot" data-
more than two decades why can't we make knowledge bases captured in commercial systems. The question is:
systems for a reasonable price, even though data and how are we going to evolve from where we are to the
communication, the basic ingredients of knowledge, database systems of the information age? What are the
are as inexpensive as integrated circuits and boards? A promising approaches? What are the right directions to
2000-year-old quotation from Aristotle gives a hint: pursue? In dealing with trends and predictions, one
"The whole thing is more than the sum of its parts." It always runs a risk of missing an important approach.
seems that we've got the parts and we can sum them up, However, the risks of not making such predictions is
December 1986 0018-9162/86/120040007S01.00 © 1986 IEEE 7
greater. It blocks imaginative concepts utes the database, its accessing, and the
from being pursued. Ten years ago, when application processing between a main-
some of us were advocating Al techniques frame and a large number of worksta-
for better information processing, some tions. It uses a set of bindings between
considered the idea as "illusionary" and downloaded and mainframe data objects
others as "prophetic." Lets hope that our to incrementally maintain the workstation
visions will materialize again. subsets of the database.
In the third article Leo Mark and I pre-
sent a multilevel architecture that allows
In this issue data to be modeled and managed by Nick Roussopoulos is an associate professor
This special issue is focusing on novel meta-data. and associate chairman for research and ad-
database system architectures to support Gio Wiederhold, in the fourth article, ministration of the computer science depart-
discusses an architecture that stores and ment at the University of Maryland. In 1985, he
information engineering. Different data- was appointed to the Space Science Board
bases using a variety of forms and special- manipulates engineering information Committee on Data Management and Com-
purpose processing have to be linked. objects. putation, an Academy of Science subcommit-
Instead of a cumbersome monolithic ar- The fifth article, by Stavros Christo- tee advising NASA. He was the General Chair-
chitecture that encompasses many inde- doulakis and Christos Faloutsos, presents man of the ACM International Conference on
pendently constructed databases and an architecture of a system that efficiently Data Management 1986. He has also organized
processors, flexible architectures provide stores and accesses multimedia objects. and chaired a series of workshops for the
a cooperative environment that is easier It is my hope that the articles of this VHSIC Engineering Information System pro-
and more economical to evolve. Clearly, issue will help the reader understand the gram. He also serves on the editorial board of
we cover only a subset of the future direc- the International Journal of Decision Support
principles behind the techniques pre- Systems.
tions in databases. We hope, however, sented. This is important because, if noth- Roussopoulos received his BS from Univer-
that other future issues will deal with the ing else, these principles have a chance of sity of Athens and his MS and PhD from the
others. enduring the evolution of engineering. D University of Toronto. He is a member of the
In the first article, Witold Litwin and ACM and the Computer Society of the IEEE.
Abdelaziz Abdellatif discuss the issue of
interoperability of multiple databases in a Acknowledgments
I would like to thank all those who con-
loosely coupled architecture that allows tributed to this special issue. For their support
access and manipulation of the databases. special thanks are due to the editor-in-chief of
The second article, by Hyunchul Kang Computer, Michael Mulder, the Editorial Readers may write to Roussopoulos at the
and myself, describes a tightly coupled Board, and one of its members, Herbert Weber, Dept. of Computer Science, University of
database system architecture that distrib- who was designated to oversee this issue. Maryland, College Park, MD 20742.
Reader Service Number 1
