The database race

The role of a database within an enterprise has evolved dramatically.Databases and tools are now more important than ever to the end user andreseller, says Sean Hallahan.

The world of databases has undergone a vast number of changes in the last 30 years, along with a number of attempts to define exactly what a database is. From the network data model of the sixties through the relational database theories of the eighties, we have jumped to today's object, semantic and parallel databases - the so-called post-relational ones. Everything is changing. While databases started out as just another application in the mainframe world they have rapidly assumed a critical role in the running of many businesses.

Database suppliers come from a variety of backgrounds. Software AG, Cincom and Computer Associates all began life in the IBM mainframe environment. Oracle and Ingres concentrated on the Digital VAX systems while the birth of the PC brought forth a number of contenders including Gupta, Ashton-Tate (now a part of Borland) and Borland itself. And there were a number of others which recognised the growing importance of Unix, such as Informix, Unify and Sybase, and directed their efforts toward that operating system.

But all these companies are third party suppliers which have to compete with offerings from the hardware manufacturers.

IBM's flagship product, DB2, was originally launched for the MVS mainframe environment in 1983 but by the early nineties the company had refined the product for the AS/400 and RS/6000. In 1995, versions were released for Hewlett Packard's HP/UX and Sun's Solaris operating systems. There is also the promise of a cutdown version called DB2 Ultralight which has been designed to compete with Microsoft Access and Borland Paradox in the OS/2 and Windows market.

IBM originally outgunned its third party database rivals when it launched the AS/400 in 1988, incorporating an integrated relational database built into the operating system and the architecture. One of the major benefits of this integration was that database disk allocation could be handled by the operating system. Some of the converts to the AS/400 from the 370 mainframe were able to reduce the amount of storage required by as much as 25 per cent, which at the time when disk storage was the most expensive hardware component of mainframes meant considerable savings for corporate customers.

Digital also integrated a relational offering, Rdb, with its VAX VMS machines although with less success than IBM. First, the company was late into the relational market for VMS behind Oracle and Sybase when it launched Rdb in 1984. Second, it never had a coherent pricing structure, as it was going through phases of supplying Rdb without additional charge and then offering it as a separate product. Finally, in 1994, Digital, hit hard by financial troubles, sold Rdb and its development environment to rival Oracle.

Over the last three years there has been a tendency for database suppliers to coalesce. Oracle's acquisition of Rdb came at the same time as Computer Associates and Borland beefed up their offerings. Acquisition-hungry CA has taken over a number of companies and products, including Cullinet, which owned a relational product originally developed for the VAX called DB (now CA-DB), and IDMS and Datacom from ADR. But the most recent and important acquisition was that of Ingres, with which Oracle carved out an early niche.

Borland currently has three offerings: dBase and Interbase, which it acquired when it took over Ashton-Tate, and Paradox. The main rival to Borland and a number of other vendors is Microsoft's Access and, more importantly for the future, the Microsoft SQL Server. The latter is the result of a joint venture between Microsoft, Ashton-Tate and Sybase, which effectively fell apart two years ago.

Traditionally, value added resellers have made their margins at the lower end of the market, although companies like CA still put some higher end products through the channel. dBase, which dominated the PC market in the mid-eighties, has now been eclipsed by Access, according to Robin Bloor, chairman of Bloor Research which has produced a number of reports on the database market and technology.

According to Bloor, there are a number of reasons why the bottom has dropped out of the low end database market. First, the majority of low end users have standardised on one product and are unlikely to switch mid-stream. Second, the majority of corporate users who originally standardised on one PC product are now looking for an enterprise-wide solution within a client/ server environment with links to larger systems. Finally, many of the smaller customers do not shop around for a database: they go into Dixons or one of the other major chains and base their buying judgement on cost and ease of use.