Well, it has been quite an eventful quarter in the database market. During the past quarter we have seen a major acquisition, updated market share estimates, new DBMS releases from IBM and Oracle,
and significant pricing moves from Oracle Corporation. In this installment of The Database Report we will analyze each of these events and place them into perspective in terms of their impact on
the overall market for database management systems.
The first significant piece of news is that Dataquest reported that the overall database market grew to $8.8 billion last year. This represents a 10% increase over the previous year. Clearly, the
database market is a strong one given that such robust growth occurred during a major economic slowdown.
Informix – A Casualty of the Database Wars
Obviously, the most important event of the past quarter is IBM’s imminent acquisition of the Informix database business. In late April 2001 IBM agreed to acquire the DBMS assets of Informix
Software in a cash transaction valued at $1 billion. Informix Software is the “new” company that consists of the database products of Informix Corporation. In August 2000 Informix split its
business into two independent operating companies: Informix Software, focusing exclusively on database software, and Ascential Software, focusing on the company’s information management
applications. Following IBM’s purchases of Informix’s database business, which IBM expects to close in the third quarter of 2001, Ascential (based in Westboro, MA) will become a wholly
independent company.
The deal should proceed smoothly with no barriers or anti-trust measures anticipated. Informix shareholders have already approved IBM’s acquisition and in early June Informix and IBM announced
that they had received early termination of the Hart-Scott-Rodino waiting period regarding IBM’s acquisition of Informix’s database business.
But Why?
Why would IBM acquire a struggling DBMS company like Informix? IBM has achieved an enviable growth rate and a significant share of the DBMS market with DB2. In a recent earnings presentation IBM
trumpeted about its 36% year to year growth in DB2 revenue. Likewise, DB2’s technology is first rate. With DB2, IBM delivers what is arguably the best relational optimization technology, coupled
with stellar performance, solid web deployment capabilities, and support for large objects and extensible data types. The DB2 DBMS technology is world class and IBM’s stands behind that technology
with very good support.
Indeed, a recent survey of database users published in Information Week ranked IBM and DB2 first in terms of customer referrals, service and responsiveness, industry expertise, reliability and
availability, scalability and VLDB, service-level agreements, and strategic advice. So why does IBM need Informix?
Well, there are a several reasons. First of all, the acquisition immediately gives IBM a larger market share on Unix platforms. According to some industry analysts IBM and Oracle each own about 30%
of the overall DBMS market. But, a lot of IBM’s market share is on the mainframe and proprietary platforms, whereas Oracle’s market share is predominantly on the all-important Unix platform.
Indeed, Informix actually has a larger presence on Unix platforms than IBM. Unix is perceived as the primary growth platform for database servers and IBM is intent on increasing its presence on
Unix specifically to combat Oracle.
So, Unix market share is the number one answer to the question “why?” But that is not the only answer. Other areas that IBM can exploit include delivering additional DB2 extenders using
Informix’s market-leading DataBlades for supporting multimedia data and enhancing IBM’s OLAP and data warehousing capabilities using analytical functions from the Informix Redbrick warehouse
product.
Another thing IBM gets out of this acquisition is some good DBMS development talent that can be used to augment their current excellent DB2 development team. Of course, that assumes that the
Informix developers stick around.
But is all of this worth $1 billion? Well, I guess it all depends on IBM’s perception of the value of these items. But it could be a big mistake. Read on to discover why.
DB2 – Still IBM’s Strategic DBMS
Make no mistake, DB2 remains IBM’s strategic DBMS. IBM has invested too much into the DB2 product and brand to abandon it. The Informix database technology will be used to augment the DB2 product.
Regardless of their current public position, IBM will be developing a migration strategy for Informix customers to convert to DB2. Of course, currently IBM is publicly stating that they will
continue to provide on-going development and support for the Informix technology and customer base. For the details on IBM’s current position vis-à-vis Informix check out the Informix and
IBM have published a white paper on the future directions of Informix products that can be downloaded at http://www.informix.com/informix/blue/future_directions.htm. IBM’s public position regarding Informix is understandable because they do not want to lose Informix customers to Oracle
or Microsoft, but it is not in IBM’s continuing best interest to support more than one DBMS. So the long-term existence of Informix as a separate, independent DBMS is not in the cards.
At any rate, regardless of what IBM says about Informix, DB2 will continue to be their strategic DBMS. Interesting Steve Mills, IBM senior vice president and group executive, was quoted as saying
“Informix customers can be confident that their investments in existing Informix products are secure.” But time will tell – the “security” of which IBM speaks most likely will be found by
migrating to DB2 instead of staying with Informix.
Over time, Informix users will be forced to migrate and IBM’s job will be to convince them to choose DB2. IBM must do this without alienating the Informix users. But remember that most of those
users have developed a loyalty to Informix over the years. And Oracle won’t just concede the Informix installed base to IBM, they will try to convince Informix users to choose Oracle instead of
DB2. The Oracle message most likely will be “Why choose to use the technology of the devil that destroyed your DBMS platform of choice? Choose an independent DBMS instead. In fact, choose the
market leading DBMS product – Oracle.”
But in the upcoming Informix conversion wars IBM will get the upper hand because they can incorporate Informix technology into DB2 to make the migration path easier. And if IBM supports a dual DBMS
strategy, at least for awhile, it is likely to help IBM win the loyalty of the Informix customer base. Furthermore, many Informix users are wary of Oracle’s pricing practices and IBM will probably
offer better terms and conditions than Oracle.
However, IBM has a long, hard job ahead of them in terms of integrating the Informix technology into DB2. Integration of existing technologies is one of the most difficult development tasks
possible. Many companies have spent a lot of money attempting to integrate acquired technology only to fail. And even if IBM succeeds it will take quite some time to merge all of the Informix
technology into DB2. So Informix and DB2 users will have a long period of waiting for integration results. Oh, IBM will have some quick successes – most likely adopting DataBlades – but long-term
integration is a more thorny issue.
More Questions Than Answers
With the acquisition of Informix IBM has created some potential problems for itself. How happy are Informix users going to be if they are forced to migrate to DB2? Informix developers are rightly
proud of the performance of their DBMS. Informix built its reputation as a high performance DBMS. Will Informix developers change their allegiance from Informix to DB2, and if so, how quickly? If
IBM does not force Informix users to migrate to DB2 how long will it take before the Informix installed base will have to migrate anyway because IBM has neglected it as compared to DB2 (as well as
other competitive DBMS products)? If IBM neglects to keep Informix up to date how will Informix users react and will they want to do business with IBM?
For Informix users, this acquisition derails the work Informix was doing on the Arrowhead project, which is essentially the consolidation and integration of Informix’s many different DBMS products
into a single DBMS. Keep that in mind – there are many different database assets at Informix, not just a single DBMS. How IBM deals with each of these database technologies will be interesting.
For current DB2 users the good news is the infusion of new database technology into IBM. But that might be the bad news too. IBM has been doing a great job at focusing on a single DB2 and keeping
it up-to-date in terms of performance, availability, and functionality. As IBM attempts to absorb the Informix technology into DB2, IBM’s database functionality track record might slip. It will be
difficult for IBM to keep up the pace of adding new features while at the same time integrating Informix into DB2.
Another positive aspect of the acquisition to current DB2 users is that IBM can use the acquired expertise to support additional operating system platforms for DB2. IBM supports its own platforms
(OS/390 and z/OS, AS/400, VSE and VM, AIX, and OS/2) very well. IBM also offers DB2 on a few other server platforms, most notably Windows 2000 and NT, Linux, and Solaris. But IBM has lacked support
for most of the less popular Unix brands (Digital, Data General, SCO), whereas Informix embraced more of these minor Unix platforms.
From the perspective of Oracle, this acquisition is probably being viewed as a mixed blessing. Of course, IBM has a much bigger claim on the DBMS market now, but most folks will be skeptical of any
claims IBM makes until Informix users begin to migrate. On the plus side for Oracle, though, is the distraction factor. Anything IBM does that causes them to take their eye off of the DB2 ball
should please Larry Ellison and the folks in Redwood Shores. And digesting Informix will be a major, ongoing distraction for IBM.
To summarize, IBM has bought some very good technology from Informix. But IBM has always had very good technology so they did not really need Informix for its technology? The strategy for Informix
and DB2 has been similar, at least with regard to their respective support for extensible data types. But if there are similarities, what exactly is the benefit of merging the two products (other
than buying a Unix installed base for DB2). The bottom line is that acquiring Informix helps with IBM’s BI/DW capabilities and it buys IBM some market share and some development talent. But is
that really worth $1 billion? That is the billion dollar question for IBM.
And What of Sybase?
Now that IBM is acquiring the database assets of Informix Software the DBMS landscape is narrowing. It is Oracle and IBM in a neck and neck race for leadership, with Microsoft continually
increasing its share of the DBMS market for Windows platforms. Oh sure, there are other DBMS products out there, but none that are really in a position to challenge the big three of IBM, Oracle,
and Microsoft. NCR has a nice niche carved out for themselves with Teradata and some very large data warehousing customers. Software AG has some devoted Adabas customers and Computer Associates
continues to milk some revenue from Ingres. But none of those three are challenging the big three. Which brings us to Sybase.
In the 1980s Sybase built SQL Server, a very innovative and technologically forward DBMS. At one point Sybase was one of the leading providers of DBMS software for Unix platforms. But Sybase came
upon some hard times in the 1990s. The biggest problem was the roll out of SQL Server System X, which had some severe performance problems. But Sybase faced many other problems including:
- a lack of focus when it acquired Powerbuilder (a leading client/server application development platform at the time);
- increased competition from industry juggernauts (like Oracle, Microsoft, and IBM) and;
- an agreement to partner with Microsoft that provided Microsoft the code base for its current SQL Server DBMS.
Sybase even had to rename its DBMS from SQL Server to Adaptive Server a few years ago to avoid confusion with Microsoft SQL Server. This is ironic because it is Sybase’s core technology that is at
the foundation of Microsoft SQL Server.
Today Sybase is the fourth place contender in a three party contest. Oh, there are still some pockets of the market where Sybase is strong, most notably in the financial industry. But few new
accounts are choosing Sybase over IBM, Oracle, and Microsoft. Yes, Sybase will win the occasional contract, but it is rare for a shop that does not already own and use Sybase to choose it for a new
project.
Even though Sybase has experienced some hard times during the past few years, the past few quarters have been quite eventful for Sybase. Sybase is acquiring New Era of Networks (NASDAQ:NEON), an
application integration company, to bolster its impressive line of middleware products. But this will not help Sybase to become a significant DBMS player again.
Sybase’s Future
If Sybase has thoughts of truly competing with the big boys on their DBMS turf they are deluding themselves. IBM, Oracle, and Microsoft can afford to spend much more money on research, development,
marketing, and sales than Sybase ever could. And Sybase has lost mind share in the DBMS community, even if they still can claim to be growing. According to AMR Research, recent year-to-year
comparison of quarterly results showed 10% gains for Sybase Adaptive Server.
Even more problematic for Sybase is the marketing campaigns that IBM, for one, is directing at Sybase customers. IBM targeted Sybase and Informix customers by offering very tempting financial deals
to switch to DB2. And if you are a Sybase customer you’d have to ask yourself, “why not?” DB2 is a technology leader and a market leader. Sybase is neither. Of course, the cost of conversion may
be enough to dissuade Sybase’s customers for the time being (but such a conversion deal will be more compelling for Informix customers now that IBM owns Informix).
Realistically, Sybase’s future most probably will resemble Informix’s present, unless Sybase retools itself as an application development and middleware solution provider. If Sybase continues to
position itself as primarily a DBMS company it is likely to be acquired by one of the market leaders for its installed base. My prediction is that it could be a target of Microsoft in the future.
But, of course, a lot depends on what happens with Microsoft’s appeal of the Department of Justice decision splitting the company in two.
If Microsoft actually becomes two companies – one focused on operating systems and another on Microsoft’s other software product lines – the non-OS portion of Microsoft may be more inclined to
look at supporting non-Microsoft operating systems like Unix and Linux. And that would make the Sybase installed base quite attractive to the Microsoft SQL Server group. Especially since Sybase
Adaptive Server and Microsoft SQL Server are based on the same core. Two Microsofts actually may spell more trouble for Sybase than one Microsoft did!
Now, if Microsoft remains a single company it is more unlikely that they will consider acquiring Sybase. The Microsoft position of ubiquitous Windows would likely continue unchanged in such a
scenario. Of course, a single Microsoft might rethink the basic core belief of Windows everywhere, but that is unlikely. Bill Gates has never been much of a Unix fan (sarcasm).
At any rate, Sybase is not coming up win, place or show in the database wars. And that puts them out of the money. With fewer resources and eroding mind share, it is unlikely that Sybase can ever
compete effectively against Oracle, IBM, and Microsoft the way they once did in the 1980s and early 1990s. Oh, Sybase can continue to survive and support the Adaptive Server product line, perhaps
indefinitely. But Sybase will not become a DBMS heavyweight again.
Microsoft’s SQL Server Business is Booming
Paul Flessner, vice president of .NET servers at Microsoft Corporation, announced at the opening keynote of the Microsoft TechEd developer conference that SQL Server is now a $1 billion dollar
business unit within Microsoft. Adding fuel to the fire was the DBMS market share numbers announced by Dataquest showing that Microsoft had overtaken Oracle on Windows platforms. The margin by
which Microsoft leads Oracle is small, but the victory is a large one for Microsoft because Windows is its own turf. And Microsoft wants to rule the market on its turf.
CEO Steve Ballmer said Microsoft is focusing on Oracle as its main competitor and leaving the mainframe side of things to IBM. “The world is moving to PC architecture servers. The real competition
and the battle there is clearly Microsoft and Oracle,” Ballmer said. This is in contrast to Oracle where IBM is viewed as the main competition. Of course, Microsoft is kind of twisting its view of
the DBMS the market to suit its needs. Realistically speaking, in terms of RAS (reliability, availability, and scalability) DB2 and Oracle are the technology leaders.
Microsoft has no major DBMS announcements planned for the near future. The next generation of SQL Server, currently in development, is code-named Yukon. The primary focus of Yukon is likely to
center on Web services through support for XML. But do not look for a new version of SQL Server any time soon.
Which Brings Us to Oracle
Oracle received some good news from Dataquest when they announced the database market share numbers in May. Oracle extended its overall share of the market to 33.8% from 31.4% in the past year.
Oracle also gained ground on IBM, which experienced a smaller increase in market share, from 29.9% to 30.1%. And significantly, Oracle still holds a commanding lead in the $3.6 billion Unix
database market, where it owns 66.2% of that particular market. IBM follows with 14.4% of the Unix database market, after which comes Informix with 6.7%. So even if you add up IBM and Informix the
pair own only 21.1% of the market, which is still less than a third of Oracle’s Unix market share.
Oracle’s big problem these days is not market share or technology, but pricing. Many users have complained about Oracle’s adoption of the Universal Power Unit, which determines price based on the
number and speed of the processors used to run the database software. IBM and Microsoft offerings tend to be much less expensive than similar Oracle offerings. But Oracle is addressing this issue.
In June, Oracle scrapped its unpopular power unit pricing option. Users disliked power unit pricing because it penalized them by charging more for higher power processors than lower power
processors. Instead, Oracle plans to charge a flat fee designed to cut prices significantly, sometimes by as much as half of the previous price. The pricing changes are being implemented to keep
Oracle at the top of the database market as competition mounts from IBM and Microsoft. In many cases Oracle will now be able to compete not just on features, technology, and functionality, but also
with a competitive price. Oracle’s pricing is still higher than IBM’s and Microsoft’s but it is now competitive, especially when you consider that Oracle is the runaway market share leader and
people tend to pay a premium to purchase from a market leader.
Furthermore, Oracle has indicated that it plans to offer its software as a chargeable service on the web without requiring front-end license fees. Such pricing is still a few years in the future
for Oracle, though. Mark Jarvis, Oracle’s chief marketing officer, expects that “software as a service” could comprise as much as half of Oracle’s income in as little as three to four years.
Oracle’s Earnings Performance
On June 19, 2001 Oracle announced better than expected earnings for the fourth of its fiscal year. Analysts were expecting earnings to come in at $0.14 a share, when in fact Oracle reported
earnings of $0.15 a share. Oracle’s earnings for the quarter were $855 million, which is actually less than the $926 million Oracle earned in the same quarter last year. But even though the
results were somewhat mixed, toward the end of June Oracle’s stock was doing well and several financial analysts had upgraded the stock.
From a database perspective Oracle’s numbers look even better. Database license revenue was came in than expected; but Oracle’s application business did worse than expected. Oracle’s total
revenue for the quarter was $3.26 billion, essentially flat year over year. Database revenue accounted for $1.28 billion which, once again, is actually 5% less than the same quarter last year. But
some financial analysts expected a more significant drop of about 8%, so a 5% drop looked good. The applications revenue though came in at $338 million, signifying a 24% year over year drop.
Oracle Rolls a 9i
When Oracle9i was released the week of June 11, 2001, Oracle boasted that it contained over 400 new or improved features as compared to the previous version, Oracle8i. In this day and age of
rapidly changing technologies Oracle is embracing change and delivering significant new features to its user base. Highlights of Oracle9i include improved clustering abilities, some nice
self-tuning and DBA task automation features, as well as improved data movement and business intelligence capabilities.
Indeed, one of the most anticipated features of Oracle9i is better support for clustering. Clustering refers to tying multiple servers together to improve scalability and reliability. The new
clustering technology, referred to as Real Application Clusters, supplants Oracle Parallel Server (OPS), which has been part of the Oracle DBMS for awhile now. But few customers implemented OPS
because it was very difficult to setup and manage, requiring developers to customize database code to work against the underlying cluster. With the release of 9i Oracle has simplified the process
and made clustering more robust. According to Oracle, any standard application can run against its clusters. Oracle needed to improve its clustering abilities because IBM and Microsoft regularly
touted their superior clustering when compared to Oracle8i.
Another interesting aspect of Oracle9i is the combined OLAP and data mining technology dubbed Advanced Analytic Services that is built into the DBMS. Oracle’s move to include such analytical
functionality in its base product is a direct response to the OLAP and Analysis services Microsoft built into SQL Server 2000. So Oracle is playing catch up here, too.
Other New Versions
In early June IBM announced that Version 7.2 of DB2. This new version of DB2 is for Unix and Windows platforms only (that is, non-mainframe). Highlights of DB2 V7.2 include better integration with
IBM’s WebSphere application server, improved e-business support with SOAP and UDDI capabilities, and like Oracle9i IBM has introduced some self-managing elements to DB2, labelled SMART
(Self-Managing And Resource Tuning).
Additionally, on June 25, Sybase announced Adaptive Server Enterprise version 12.5. We discussed this version in The Database Report last quarter, so refer to that column for the highlights of ASE
12.5.
Summary
We have reached the end of another Database Report. And the conclusion is that the market for DBMS products and services is prosperous and active. Modern applications rely on DBMS software for
persistent data storage. As such, organizations will rely on the major DBMS vendors to supply database software for their data processing and storage needs.
Thanks for reading – and we’ll see you next quarter to review what is going on in the database wars.
