0007-0009

Previous Table of Contents Next

Page 7

DB2 itself, IBM's flagship relational database, beat Oracle in its only arena of competition, which was IBM's MVS mainframe environment. The fact that Oracle could offer a database on a platform so remote from UNIX or VMS spoke strongly of its portability. As the world moved off these large legacy systems created by IBM, Oracle was more than happy to secede this sacred ground.

Once Oracle moved to version 6.0 of the database, which was a reworked kernel that better handled multiple user writes to disk, the speed issue between Sybase and Oracle evaporated. Oracle's SQL*Forms improved and gained even a larger share of the market. Oracle naturally followed IBM and DB2 database syntax and the basic form and use of language precompilers; therefore, Oracle sought the common standard and strove for the image of stability.

As the wars continued in the 1990s, Oracle was perceived as the safest choice by far for the average customer. You knew the company had a good product and would be around for a long time; it was no longer a startup but a billion-dollar firm. Version 7.0 only enhanced Oracle's position and closed any technical gaps with its competitors in terms of features, such as database triggers, which Sybase fans had touted for years . Oracle also extended the PL/SQL language as the basis for Oracle database development.

Oracle Today and Tomorrow

Today Oracle is confronted with three new developments in computer science. It appears the future of the fastest database hardware is the massively parallel machines. These machines run multiple CPUs in parallel, more like the human brain. The database running on these machines must use parallel processing. Oracle, which is well positioned in this arena, anticipated this technology earlier than its competitors. In 1991, Oracle reached 1,000 tps (transactions per second) on a parallel computing machine. Oracle's RDBMS was the first relational database to run on a massively parallel computer.

The second issue that Oracle must tackle is the new popularity of the Internet and the World Wide Web. Although the Internet existed before Oracle did, it has only recently become popular. Oracle now must build products that allow companies to do business on the Web. With the Oracle Web Server and the Electronic Commerce Server, Oracle also appears to lead the database vendor pack in this second arena of battle.

Today Oracle has turned away from the other database vendors and looked toward the software giant Microsoft, which ironically also began as a startup based on an IBM technology. This is the third arena where Oracle's Larry Ellison is fightingthe attempt to break the Microsoft software juggernaut . His answer is the Network Computer Architecture (NCA) and the Common Object Request Broker Architecture (CORBA) industry standard. Network computers have a more open interface than Microsoft Windows and Windows NT products and are a cheaper way to build intranets . The network computer also moves more processing to central serverswhich, of course, is Oracle's area of expertise.

Page 8

If Oracle and other vendors such as Sun with its Java language can convince the world to move to NCA and CORBA, Oracle will probably become the largest software vendor in the world. Oracle must balance this goal, which entails moving into uncharted territory for a database company, with the necessity to constantly revolutionize and improve its existing database server technology and remain in possession of the best database server in the market.

At the end of May 1997, Oracle reported earnings of $5.7 billion, which was an increase of 35% from last year, a significant jump. Oracle accounts for 29.8% of the market share (followed by IBM's 25.6%), which makes it the leading database vendor in the world. No one can predict for certain the future of the company, but it is safe to say that its prospects look bright.

Oracle's New Baby: Network Computing Architecture

Today, Oracle's CEO Larry Ellison is not spending most of his time talking about relational databases; instead, he champions the NCA (Network Computing Architecture). The network computer is a vision of a less-expensive system on the desktop that simply downloads applications and applets from a main server. In a corporate setting, this would help alleviate software inventory and networking nightmares. You would not need to upgrade spreadsheets, word processors, and other applications on thousands of PCs every few months; you could simply change them on the servers.

Using a computer would be more like using a telephone. "All mature networks like TV and telephone have the same modelsimple appliances with sophisticated and powerful networks," Ellison said in a May 1997 Oracle press release.

"Users shouldn't have to worry about the underlying technology," he continued. "Developers shouldn't be left in the cold by a proprietary standards battle. And corporate customers shouldn't be stranded with old technologies that don't work together."

The Goals of the NCA

The vision of network computing hopes to improve upon the following six areas:

• Portability: Today it is difficult to port an application among Windows 95, UNIX, and the Macintosh. With a Java-based development standard, portability of Java applets is not part of the application code but instead is the responsibility of the Java virtual machine for each piece of hardware. Once a developer develops an applet or application in Java, the code is already portable among Apple, Windows 95, and many UNIX machines because they all have Java virtual machines.

Page 9

• Salability: An application that can talk to different kinds of servers with different host operating systems can easily move from a smaller machine to a larger, faster machine as long as both machines run Java.
• Interoperability: With all new NC applications, interfaces between applications and rules for interfacing are already defined by Java standards and the new CORBA standards. You won't have a C program trying to speak to a COBOL program that gets data from CICS screens; all code is written using the common Java language. This does not exclude other languages but requires a Java "wrapper" around everything, thus making it easier for different vendor products to communicate with each other.
• Component capabilities: With Java as a standard interface, along with standards such as CORBA, software components will become standard. You won't have to rewrite subroutines to perform the same task because component libraries will thrive under the CORBA standards.
• Data extensibility: With standards such as CORBA, you can pass more complex objects from application to application. In the past, the standard datatypes were numbers , strings, and dates. Now a standard object-request protocol can handle multimedia objects, sound, and more complex business objects such as spreadsheets or engineering diagrams.
• Migration: With the NC on the desktop, migration to newer servers becomes transparent to the user and almost transparent to the application programmer. Because Java is an interpreted language, any hardware vendor with a Java virtual machine can run Java applications written on other platforms.
  Also, with a standard NCA, you can easily integrate new advances in the graphics, sound, or operation of network computers with existing servers.

What Is the Difference Between the Network Computer and What We Have Today?

The goal of the network computer (NC) is to replace today's desktop, which is dominated by Microsoft. It is a technology directed toward corporate or educational environments. When you first glance into the window of the network computer mindset, you might immediately think this architecture is a throwback to the days of mainframes with dumb terminals. From the simplest perspective, NCA is a central server with all the necessary applicationssuch as word processing, spreadsheets, e-mail, and database capabilitiesthat feeds to the NC itself, which is a computer that simply requests information from the server (see Figure 1.3).

Previous Table of Contents Next