Turning The Tables On Oracle
Sydney Morning Herald
Tuesday June 1, 1999
MOST computer systems nowadays have at their core a database management system (DBMS). Over the last 15 years or so, the DBMS market has become the most keenly contested in the computer industry, because of the software's key position at the heart of the organisation's IT infrastructure.
DBMS software sorts data into rows and columns (called tables) and makes it far easier to manage. In the early days of computing, every program had to work out its own data storage method. A DBMS handles this important task, interfacing with the applications software to seamlessly store and manipulate the vast amounts of data typically used in today's computer systems.
The biggest advance in DBMS software in the past decade has been the almost universal adoption of the relational DBMS model, which enables any number of two-dimensional data tables to be linked to each other to build data models of great complexity.
There are a half dozen or so major players in the DBMS market. Largest by far is Oracle, which is increasing its lead each year. There is a general trend in the IT industry that the big players get even bigger, and the smaller players fall behind.
Oracle's biggest rivals in the 1990s have been Sybase and Informix. As recently as five years ago they were often mentioned in the same breath as Oracle, but they have fallen far behind and are now second-rung players.
Both faltered badly a couple of years back. Sybase hit a wall with its technology, shipping a major new release before it was ready, and lost a lot of money before it got back on track. It fixed its technology problems, but the financial fallout took a while longer to repair, and it recovered by cutting costs, and development, as much as by increasing revenues.
Informix had an even tougher time. It was the rising star when it purchased Illustra, a company with a new type of technology called object relational, but it all went horribly wrong when it was caught falsifying revenue figures to add to its illusion of growth. Those responsible "are no longer with the company", and it has concentrated on consolidating its user base more than on expanding its activities.
There was another player, Ingres, which fell on hard times and went through a few changes of ownership before being bought by perennial acquisition merchant Computer Associates a few years ago. Ingres was always very strong in Australia, but it has been weak in the United States, and a decreasing number of applications software suppliers are supporting it, ensuring its decline and prob- able eventual death.
There are many other DBMSs. Progress is the most important of them. Many people don't know about it, but it has achieved a reasonable market share through being embedded within many popular applications, such as QAD's MFG/PRO manufacturing software.
Oracle no longer regards these pretenders as major threats. They remain competitors, but none have the ability to challenge Oracle's momentum. Ask anyone at Oracle who they see as the company's biggest DBMS threat, and they will invariably answer "Microsoft".
Microsoft's DBMS is called SQL/Server (invariably pronounced "sequel server"). This fast-growing DBMS was originally developed in conjunction with Sybase, but Microsoft decided to go its own way about five years ago.
Unlike its rivals, which run on a whole range of operating systems, SQL/Server runs only on Windows NT (now renamed Windows 2000). SQL/Server's fortunes are largely therefore tied to those of Microsoft's flagship operating system, and have improved at around the same substantial rate.
As Windows NT/2000 improves in performance, and as it is being increasingly considered for corporate applications, so SQL/Server is finding its way into more and more significant applications. The number of organisations in Australia using it for significant systems doubled last year.
The other major player in the DBMS arena is IBM, which markets its products under the Official Secrets Act. IBM essentially invented the relational DBMS. At least, an IBM employee named Ted Codd did, but IBM didn't know what it had and the good Dr Codd moved on to become an iconic figure in the DBMS world.
IBM has for many years had a DBMS on its mainframe computers called DB2, a very good and very powerful DBMS which has become the system of choice for most large mainframe users.
Contrary to popular belief, there are still plenty of these - mainframes never died. It is simply not possible to run large transaction processing systems like an ATM network or an airline reservation system on any other type of machine.
Following the success of DB2 on the mainframe, and the rampant growth of DBMSs on midrange Unix machines, IBM thought it would compete against the Oracles and Sybases and Informixes of the world with a new DBMS. In a stroke of marketing madness, IBM decided to call its new DBMS DB2, the same name as that used for its mainframe product. It hoped that some of the success of the older DB2 would rub off on the younger product, but the opposite happened.
Hardly anybody bought DB2 for Unix (also known as DB2/6000, after IBM's RS/6000 Unix box), because they identified it with the mainframe product, and thought that it was therefore old technology. IBM should have called it DB3 or DB4 or DBX, or even Frank or Eric, to differentiate it from the original DB2.
A couple of years ago I visited IBM's DBMS development labs in Toronto, and in Santa Theresa near San Jose. I got the full two-day rundown on the various DB2s (there is another for the AS/400 called DB2/400). IBM was very good to me, and even looked after my mother with shopping trips and the like. But I still can't understand IBM's DBMS strategy.
Just last week IBM announced a major new release of DB2 for Unix, which is now called Universal DB2 because it also runs on Windows NT. It has fluffed yet another opportunity to rename this ill-fated product. UDB might have worked, but it's probably too late now. Oracle has won, and SQL/Server is the big threat.
DB2, or UDB, or Eric, or what ever it is called, has missed the boat.
geepee@ozemail.com.au
© 1999 Sydney Morning Herald