The Database Report – October 2001

As I write this edition of The Database Report it is just one week after the terrorist attacks on the World Trade Center Towers in New York City and the Pentagon in
Washington, DC. With the prospect of real war hanging over the world, it seems wrong to refer to the events impacting the DBMS market as the “database wars” as this column so frequently does.
Instead, let’s just analyze the business and marketing drivers impacting the market and usage of Database Management Systems in the quarter ending in September 2001.

As usual, we will discuss the events that impact the big three DBMS vendors: Oracle, IBM, and Microsoft. But this quarter we will also examine the growing market for open source DBMS products.
Let’s start out with the good news at Microsoft.

No Government Breakup for Microsoft

The biggest news this quarter for Microsoft fans is the reversal of the decision to break the software goliath into two separate companies. As it stands today there is no way that Microsoft will
not continue as a single company once all of the legal activity is completed.

After the court reversed the initial decision, the Justice Department later declared that it will not seek Microsoft’s break up in new court hearings. Instead it is looking more and more like the
Department of Justice (DoJ) and Microsoft will reach a compromise agreement. If calmer heads had prevailed, both at Microsoft and in the government, this is what would have occurred more than a
year ago. But perhaps all of this legal wrangling was required to get both sides to the table to hammer out a “reasonable” resolution.

The likely outcome will be a series of restrictions on Microsoft’s conduct that includes forcing Microsoft to disclose its APIs, provide versions of Windows without Microsoft middleware bundled
in, and stop using pricing and access to technical information to coerce PC makers into onerous contracts. From the DBMS perspective, though, any remedies that may be imposed on Microsoft are
unlikely to impact SQL Server users, other than perhaps an impact to the operating system on which it is used. Most of the issues that need to be resolved focus on desktop operating systems (and
the web browser). Of course, there are desktop editions of SQL Server, but users predominantly run SQL Server on a server.

But the next phase of the Microsoft antitrust case could be as long as the first phase if no settlement is reached. Both sides, Microsoft and the DoJ, are preparing to their strategies for deciding
what remedies to place on Microsoft – and initial reaction on recent court filings show both sides gearing up for a fight!

Even if an agreement is reached, keep an eye on what happens with Windows XP. Microsoft’s new registration and activation procedures are likely to draw the ire of users and might perk up
additional legal activity against Microsoft. Once again, though, the good news for Microsoft SQL Server users is that the company may be able to focus more energy on product delivery and innovation
after it clears up its existing legal hassles.

IBM Reasserts Its Commitment to Informix

Last quarter we evaluated the IBM acquisition of the database-related assets of Informix Software and came to a mixed conclusion. IBM must continue to lead with DB2 as its enterprise database
management solution of choice; but Informix customers are not likely to want to convert to DB2, at least not immediately. So IBM is stuck with having multiple relational DBMS product lines.

This past quarter IBM outlined its plans for delivering new versions of DBMS products across the Informix product line. IBM plans for Informix Dynamic Server 9.3 to ship by October 1. IBM is
touting the new version’s performance enhancements and enterprise object-replication capabilities. A new version of the Informix XPS database for building data warehouses is scheduled for
availability in early 2002, or perhaps even by the end of 2001. And version 6.2 of Informix’s Red Brick data warehouse is slated for Spring 2002 availability; among its new features is advanced
data mining capabilities.

IBM has committed to producing at least two releases of each of the Informix products. Furthermore, IBM has stated its intention to deliver additional upgrades to the Informix product line within
18 months.

More importantly, IBM is planning to build Informix features, such as time series and spatial data, into DB2. This is the real future of the Informix technology – as components used to augment DB2
functionality. IBM cannot indefinitely support both DB2 and the multiple Informix versions and product lines. Sooner or later the eventual status of Informix will be as a legacy DBMS with little
new functionality being added – because IBM will, nay must, focus on DB2 functionality first and foremost.

In an interview published in The Register (a technology newspaper based in the UK), Marc Dupaquier, IBM’s worldwide VP for data management solutions sales provided some additional details on the
Informix acquisition. He stated that the IBM and Informix sales teams were merged on July 2nd, right after the deal finalized. Further, Dupaquier asserted that IBM did not just force the Informix
people to report to IBM people. Another interesting detail supplied by Dupaquier is that IBM lost only three Informix employees since the acquisition – quite a good retention rate that speaks well
for IBM’s ability to digest Informix in the long term.

The most troubling aspect of this acquisition from the perspective of long-time IBM DB2 customers continues to be the lack of a clear and strategic plan for relational database management systems
at IBM. What is the road map for integrating IBM’s portfolio of relational DBMS products? So far, IBM’s strategy for the Informix product line seems to be to run it the same as Informix did
before its acquisition. Which demands the question – “What makes IBM think it can manage a confused and unfocused product line consisting of multiple versions of the Informix database engine any
better than an independent Informix could?” This question has never been answered well by IBM.

Obviously, IBM will continue to focus its marketing and sales efforts on DB2. So why does it not clearly elucidate the benefit of the Informix acquisition to its DB2 customers? Instead IBM appears
to be too afraid of disenfranchising Informix customers – a laudable goal – instead of touting the benefits of the Informix technology to its DB2 customers – a strategic blunder if IBM continues
along this path very long. Face it, the Informix DBMS was slowly dying when IBM bought it and it will continue to lose market share. IDC reports that Informix DBMS revenues declined from $701,000
in 1999 to $571,000 in 2000 (representing a 4.6% share of the overall DBMS marketplace).

So the good news out of IBM this quarter is more for its Informix customers than for its DB2 customers. But IBM did have a few newsworthy events this past quarter. For example, at IBM’s Software
Solutions Conference in San Francisco, Janet Perna, general manager of IBM’s data management solutions division, discussed the integration of DB2 and IBM’s WebSphere application server product
family. IBM asserts that the combination of DB2 and WebSphere will provide the ability to better perform data analytics to gain insight into data trends.

Additionally, IBM continues to announce additional tools for managing and tuning DB2 and IMS databases. Quite obviously, IBM is ramping up to compete with the major DBA tool vendors like BMC
Software and Computer Associates. But the major inhibitor to IBM’s success in the DBA tools market is its inability to offer heterogeneous management tools. In other words, IBM only offers tools
that manage IBM databases, but most of its customers have a combination of databases implemented using Oracle and SQL Server, in addition to DB2 (and IMS).

Other News From the Major DBMS Vendors

On September 13, in the aftermath of the September 11th terrorist attack on the United States, Oracle Corporation announced its fiscal first quarter results without comment or elaboration. Details
of the announcement can be found at http://www.oracle.com/tellmemore/?892540. Generally speaking, the news was good. Consensus
earnings estimates were.08 per share, which Oracle beat by a penny. Revenue was slightly lower than expected though, $2.24 billion instead of $2.35 billion. Research and development spending was
basically flat.

Then, on September 17, Oracle issued a warning about its fiscal second quarter revenue projections. Even though Oracle reported that a slowdown in revenue growth from new software licenses showed
signs of bottoming out during its first fiscal quarter, it asserted that the terrorist attacks on New York and Washington, DC could hurt the company’s financial performance. I think this is the
kind of announcement that many technology companies are likely to make in the upcoming weeks. It is not a situation that applies only to Oracle, but across the board. Airline, financial and
insurance industries could delay software orders because of the terrorist attacks.

Indeed, a few days later, on September 19, Sybase issued a similar warning. John Chen, Sybase CEO, stated that Sybase expects business volumes to be radically affected for the balance of the
calendar year. But Chen did state that he expected Sybase to “bounce back” after that.

Earlier in the quarter, toward the end of August, Sybase tried to re-assert itself as a DBMS player by significantly changing in its pricing policy. Sybase moved toward a per-processor-based
licensing model, with a price of $14K per processor (for all new configurations in North America). This is roughly 30%-50% less than Oracle’s pricing. Such a move is smart for Sybase because it
may help to reduce the number of defections from its Adaptive Server Enterprise offering. But few of Sybase’s long-term customers are leaving. Instead, many Sybase shops are deploying DB2, Oracle,
or SQL Server for new development projects, instead of buying more Sybase product. Perhaps a pricing effort can reverse this trend. But Sybase still has a rough road ahead of it as the number four
player in a three-company race.

Open Source DBMS

The media circus surrounding Linux and its rapid acceptance as a Microsoft-alternative has enlivened the open source community. The term “open source” refers to software that users are free to
run, copy, distribute, study, change and improve. Often “open source” is interpreted to mean free software. This is understandable, but incorrect. With open source, the concept of free is closer
to liberty than to no charge. Open source software adheres to the following beliefs:

  • Users are free to run the program, for any purpose.
  • Users are free to inspect the actual source code of the program to determine how it works.
  • Users are free to modify and adapt the software to their specific needs.
  • Users are free to redistribute copies to whomever.
  • Users are free to release their code improvements to the public, so that the whole community benefits.

Linux is an open source operating system based on UNIX. Companies like Red Hat and Caldera have become quite successful by packaging and providing support for Linux. Even established companies are
embracing Linux: Corel distributes a version of Linux and has ported its Office suite to Linux and IBM has ported a version of Linux that runs on their System 390 mainframe. But operating systems
are not the only programs available as open source. There are several open source database management systems that are gaining popularity, namely MySQL and PostgreSQL.

MySQL

MySQL is a true multi-user, multi-threaded SQL database server. It was created by Swesish-based MySQL AB, and is available as a free download from http://www.mysql.com. According to the General Public License under wish MySQL is distributed you do not get access to the source code, and therefore are not able to
modify it. And you must pay for a license if you sell MySQL, charge for administering a server, or include MySQL as a component of a distribution. Furthermore, MySQL is free of charge on UNIX and
OS/2 platforms only. For Windows platforms you must purchase a software license. A free, slower version of MySQL is available for Windows platforms on a 30-day trial period. So, MySQL is close to
open source, but not exactly open source in the true sense of the term.

MySQL is a good alternative to high-cost, commercial DBMS products for small- to mid-range database servers and applications.

In July, software vendor NuSphere launched a web site called MySQL.org as a community for developers. MySQL AB was less than thrilled with the move saying it has no problem with a site devoted to
MySQL, but the name is too similar to MySQL.com. MySQL AB feels that too much confusion will arise between NuSphere’s MySQL.org and MySQL.com.

And this is not the only flare-up between MySQL’s creator and NuSphere, one of its largest distributors. Previous disagreements centered on software licenses and contracts and the companies filed
matching lawsuits against each other.

So, as an open source DBMS, MySQL has some challenges. The utopia of developing and freely sharing advancements is not quite reality in the world of MySQL.

PostgreSQL

PostgreSQL was created at the University of California, Berkeley as the second generation of University Ingres. The Ingres code was enhanced into a commercial product by Relational Technologies.
Relational Technologies became Ingres Corporation, which was purchased by ASK Corporation, which was then purchased by Computer Associates. Ingres still limps along as a commercial product marketed
by Computer Associates.

But University Ingres was enhanced at the University of California, Berkeley, too, eventually turning into the second generation Postgres. Both Ingres and Postgres were the brainchildren of Michael
Stonebraker, founder of Illustra (which was acquired by Informix, then IBM). PostgreSQL is also commonly referred to simply as Postgres and I will use the two interchangeable in this article.

The PostgreSQL DBMS is released under a Berkeley-style license, which, unlike the general public license used by Linux (for more information http://www.fsf.org/copyleft/gpl.html), allows companies to take the software and add proprietary extensions. With Linux, by contrast, companies or
individuals are required to publicly release any changes made if they choose to distribute the software.

Postgres is a fully functional object/relational database management system. PostgreSQL version 7.0 is compliant with the entry-level SQL92 standard. Full compliance with the SQL92 standard should
be available in the upcoming 7.1 release. Postgres supports multiple types of indexes, a very complete set of data types, row-level locking, triggers, constraints, and most of the features required
of a robust relational database server today.

Until recently, Great Bridge (http://www.greatbridge.com), a subsidiary of media company Landmark Communications, had plans to become the
Red Hat of the Postgres world. Red Hat was one of the first companies to package, distribute, and support Linux, and had a very successful IPO in 1999. Great Bridge was formed in early 2000 with
aspirations of leading the next wave of the open-source software – database management systems.

But alas, that was not to be. Great Bridge called it quits in early September 2001. Of course, this does not spell the demise of PostgreSQL, just its biggest supporter in the market. Of course,
other supporters exist. In early July Red Hat took up the cause and now offers a version of PostgreSQL it called Red Hat DB. Red Hat actually has a better chance of succeeding with Postgres than
Great Bridge ever did. By coupling the Linux operating system with the PostgreSQL DBMS, Red Hat will likely be able to entice more corporations to view open-source solutions as an attractive
proposition.

Yet there are several hurdles for any version of PostgreSQL to gain acceptance in the market. The first of which is the clear failure of Ingres to capture and hold a large percentage of the market.
Although Ingres is still marketed and sold by Computer Associates, it is not one of the major RDBMS players today. Those with a long memory may consider Postgres to be a commercial failure because
of Ingres’ failure to capture market share.

Additionally, Michael Stonebraker used Postgres as the basis for the Illustra object/relational database server. Illustra was the first commercial object/relational product and revolutionized the
market for storing large multimedia objects in relational databases. Informix acquired Illustra and used the Illustra technology as the basis for Informix Universal Server. Now that IBM owns this
technology some users may wish to purchase it from IBM and get IBM’s world-renowned support instead of going the open source route with PostgreSQL, even with Red Hat support. Many organization
believe that a large commercial company the size of IBM is much better suited to support and extend their DBMS.

But many of the users who embraced Linux as an alternative operating system to the many flavors of commercial UNIX and the Windows platforms offered by Microsoft may view PostgreSQL as just the
RDBMS for their Linux server. And as the market share for Linux grows, so may PostgreSQL. Only time will tell.

If you are looking for a low cost alternative to the traditional relational DBMS vendors, one of these open source or low-cost relational database servers may be just the ticket for you. But be
aware that these alternatives will usually lag behind Oracle, DB2, and Microsoft SQL Server in terms of features, functions, and performance. Even more problematic is the lack of database
administration tools like schema change managers, SQL tuning, and performance monitors. Without these tools, administration will probably be more difficult than you are used to. But if cost is an
issue, your database size is manageable, and usage is low to medium, PostgreSQL and MySQL may be viable options for you.

Linux and the Major Vendors

Of course, another alternative for open source DBMS is to use Linux – the open source operating system du jour – with a battle-tested industry leading DBMS, instead of an open source DBMS. All but
one of the leading DBMS vendors offer a version of their DBMS that runs on Linux; that is, Oracle, Sybase, and IBM. Of course, the one missing DBMS vendor is Microsoft. Microsoft SQL Server runs on
Windows NT, period. And Microsoft most likely will never develop a Linux version of SQL Server because it would eat into Microsoft’s Windows NT revenue.

And there is still one more potentially formidable player in the open source market place – SAP. At Linux World late last year SAP announced it will make its SAP DB database management system
available as open-source software under the GNU General Public License. The SAP DB DBMS is SAP’s option for organizations that do not already have a DBMS, but want to implement their enterprise
resource planning (ERP) software, R/3. As part of opening up SAP DB as open source software, SAP also agreed to host a community forum at http://www.sapdb.org to fuel open source development of SAP DB.

This is basically good news for open source DBMS proponents. SAP is a huge software vendor who can provide substantial backing to the open source DBMS movement. Furthermore, SAP has a lot to gain
if SAP DB can be used to reduce the total cost of ownership (TCO) for its R/3 customers. On the negative side, SAP DB is used by only a small percentage of SAP’s customers. Furthermore, most of
SAP’s larger customers already have an investment in an enterprise level DBMS such as DB2, Oracle, or Microsoft SQL Server. Finally, there are few, if any, management tools available for open
source DBMS products, SAP DB included. This tends to make database administration and management more costly.

But the bottom line is that the open source movement has indeed hit the DBMS marketplace. And it is likely that one or more of the open source DBMS contenders eventually will become viable for
enterprise database management.

Summary

Another quarter has come and gone. Of course, the changes to our lives have been much greater than any changes in the software market. Our hearts go out to the people affected by the tragedies of
September 11, 2001. But our lives must go on. However, before they do, we can all work together to continue our way of life by enjoying and expressing our freedom as we always did. The legacy of
these tragic events must be that the American people continue to live free as we have always done.

So, until next quarter, keep those databases up and running…

Share this post

Craig Mullins

Craig Mullins

Craig S. Mullins is a data management strategist and principal consultant for Mullins Consulting, Inc. He has three decades of experience in the field of database management, including working with DB2 for z/OS since Version 1. Craig is also an IBM Information Champion and is the author of two books: DB2 Developer’s Guide and Database Administration:The Complete Guide to Practices and Procedures. You can contact Craig via his website.