Biblio Tech
Review
Information Technology for Libraries

Database update

Register with BTR

Biblio Tech Review
2001 Issues
Tech Briefings
Product Directory
Exhibitions
Archive Issues
Archive Reviews
About BTR

Technology section

[Product DBMS]
[
Database update]

Line

Database Update                               

Database companies and technology news

Contents

Comment

The database world is beginning to shake a little as some of the established players begin to suffer with poor profit results.  Some of this is due to problems with revenues in Asia following the economic problems there and some of it is due to deeper changes in the IT world.

All suppliers have problems. Informix and Sybase have both seen r

“One factor effecting the world of databases is the strain that modern styles of application building are placing on the traditional style of server software.”

evenues fall and have taken losses in recent years.  Even Oracle have had problems as profits fell at the end of last year - although their 3rd quarter results showed a good recovery.

In the LMS (Library Management System) marketplace, there is a growing trend towards Oracle as the database of choice - see chart. Some major systems have moved to Oracle over the last few years and Oracle is now the majority supplier in the market whereas a few years ago Oracle was rare and library system vendors were either sceptical about the suitability of the relational model or preferred other more technically elegant solutions like Sybase.  Pundits are now saying that Oracle, IBM's DB/2 and Microsoft's SQL Server are the three most likely to survive long term.

Strangely no major supplier uses the IBM's DB/2 which works well on a diverse range of hardware platforms including Windows, Windows 95, Windows NT, OS/2 Warp, AIX/UNIX, Hewlett-Packard HP-UX, Sun Solaris, Apple Mac OS, MVS, VM, OS/390, and OS/400.

One factor effecting the world of databases is the strain that modern styles of application building are placing on the traditional style of server software.  Just look at the sort of computing done in libraries now to provide a service to clients.  The Internet has focussed attention on the importance of delivering object oriented data to the desktop. Users now expect an application to deliver Web pages, complete with multi-media, Java applets and other complex data. 

If the application has to deliver these objects then the database has to store them. The traditional relational database was designed to store basic data with strict rules concerning the relation of a say a title to the authors responsible for it. Now a database has to store a Web page or Java Applet along with the rules for its subsequent use.  This is what object oriented databases are designed to do and what  many relational databases are struggling to do with add-ons and re-engineering.  Oracle for instance has announced several new features within Oracle8 and the associated developer tools to handle objects.  CA have launched "Jasmine" - a pure object database which is going to further attack the weakened database vendors.

As some vendors have moved towards Oracle, DRA, with Taos, has jumped a level and gone straight to an object oriented database architecture. Libraries have to take a long view in everything that they do. How do these shifts in database technologies and business patterns affect purchase decisions?  What about the suppliers who still use proprietary technology?

The answers must be to look carefully at the DBMS used by your potential vendor or incumbent supplier and ask the questions - what if there are problems with support and long term viability of the underlying technology? And how do you handle objects within the databases structure?

Those vendors who have isolated the DBMS structure from the application will have the most convincing answer to the first question since they should be able to move easily to a new DBMS supplier - Amicus for example has already migrated from Ingres to Oracle.  Vendors using their own proprietary database have to worry more about whether they have the resources to re-engineer for objects and other industry developments as the pace of change increases.

Almost all of the main database technologies are adding some object handling features - the question is whether the greater efficiency and speed of application development with the true object databases will win out over the sheer marketing muscle of the likes of Oracle as they themselves try very hard to re-engineer their products.

Table showing database used by library systems - these figures are derived from the table of products and databases in the database technology briefing.

DBMS type or product

number of systems

Object Oriented

1

DB/2

0

SQL Server

1

Oracle

8

Informix

2

Ingres

1

Sybase

4

Nested DBMS Universe/Unidata etc.

4

Proprietary

3

Others

4

Total

29

Oracle Update

Oracle have released a whole bunch of Web related tools over the last year and are now a serious Web tools provider - any supplier that uses Oracle as a background RDBMS has a rich set of facilities to choose from when building systems for the Web and desktop.

In addition, April 15th will see the release details of version 8.1 of Oracle which will feature a Java virtual machine for executing Java triggers in the database.  This should allow for greater portability of applications across platforms and avoid the problems that all suppliers face if they support more than one operating system environment e.g. flavours of Unix and NT. Both Informix and Sybase have announced similar features.

Other Oracle announcements have included the bundling of Oracle8 with Novell Netware version 4.1. from June 1998 This will enable users to run an Oracle server without having to run an NT or Unix server.

Finally, Oracle have announced a version of Oracle (Oracle Lite) for the portable Windows CE, the PalmPilot and also Psion hand held computers - a library system in your pocket!

Market share

The latest figures from the industry watcher Datquest, show that for 1997 show that the big three have consolidated their positions with the rest of the market virtually nowhere. There will always be a market for niche DBMS which have special features not required in the main business areas, however, it is the market leaders who will drive the standards and their R&D dollars that will define the market.

Company

1997 %

1996 %

Oracle

27.5

24.9

IBM

27.2

27.2

Microsoft

14.9

12.1

Sybase

4.5

5.7

Others

25.9

30.1

Top of page