Well, the second quarter of 2002 is officially over and it was a stormy one for the DBMS marketplace. DBMS vendors and analysts slugged it out over who was number one, one vendor was accused of
price gouging, and, oh yes, some new versions and useful features were actually delivered. Let’s take a look at the turbulent second quarter that just ended.
We’re Number One! Who’s Number One?
During the past quarter both Gartner and IDC released research reports on the status of the DBMS market place. First up was Gartner in early May. Interestingly, the Gartner report showed IBM as the
overall market share leader for database management systems. This was big news because Oracle, the perennial leader, was pushed into second place. According to Gartner, IBM’s ascent to the top of
this market was due to its acquisition of Informix in 2001. IBM’s market share stands at 31.7% without Informix, but at 34.6% with Informix. Gartner’s number two in terms of market share is
Oracle, with 32%. Microsoft came it at number three with a 16.3% share of the DBMS market.
Predictably, each of these top three DBMS vendors produced a press release to spin the “research” their own way. IBM crowed about its overall leadership position, Oracle declared themselves
leader of the modern DBMS market, and Microsoft touted its leadership position on the Windows platform.
Of course, each of these positions is somewhat accurate and quite defensible. According to the Gartner numbers, IBM is indeed the market share leader – so its claims are accurate. What about Oracle
with its claim of leader of the “modern” DBMS market? Well, Oracle’s claim, too, is understandable. The IBM numbers include revenue from IMS – IBM’s hierarchical DBMS that runs on mainframe
systems. IMS is a stalwart DBMS that runs many of the most important legacy applications for thousands of the largest companies on the planet. But it is not a modern DBMS in this relational age.
Furthermore, Oracle makes the tenuous claim that mainframes in general are not modern, either. The Oracle press release, which can be found at http://www.oracle.com/corporate/press/index.html?1335580.html, states that Gartner should only be analyzing DBMS
revenue on Unix, Linux, and Windows platforms. Of course, this would help Oracle immensely since much of IBM’s revenue comes from mainframe platforms, but it would not be indicative of a true
“modern” DBMS market. Many modern database applications rely on a mainframe running DB2 as the database server. And Gartner publishes a section of their report that covers only Unix platforms –
on which Oracle has a 63% share and IBM comes in second with about 25%. So, part of Oracle’s argument seems to me to be valid, but other parts are just nonsense.
Microsoft, the third combatant here, makes a valid claim as well – that they are the leader on Windows platforms. But no one really expected otherwise, did they? The bigger surprise here would have
been if Microsoft were not the leader on Windows platforms, wouldn’t it?
Then, hot on the heals of the Gartner report, IDC released their own research – and would you believe it – it showed different results! The IDC research indicates that Oracle is the overall DBMS
market share leader with a greater than 10% market share lead over second place IBM. Microsoft, again, places third.
Why the big discrepancy between the two research reports? Why should two reports on the same market for the same time period issued within days of one another show such markedly different results?
Truthfully, there are all kinds of reasons. Here is a laundry list of them:
Analyst methodology – analysts rarely accept vendors at face value. Each analyst group has their own methodologies for interpreting and extrapolating the financial numbers
provided by the vendors.
Hidden details – the way that the vendors report their earnings is not the way that the analysts report DBMS market share. Are database tools lumped into the DBMS revenue? Is
other software combined with the DBMS revenue? How does each analyst group treat these anomalies?
Human bias – analysts are continually briefed by these DBMS vendors throughout the year. It would not be unexpected if the analysts’ opinion on the vendor’s strategy found its
way into the market share reports.
So what does all of this market share stuff really mean and whom should you believe? Well, here is what I think: Oracle is still the leading DBMS vendor and will probably continue to be for some
time. IBM is clearly picking up steam, gaining users on non mainframe platforms, and is entrenched in the number two spot. It does not seem to me that Oracle is losing past customers but that they
are not gaining new customers as fast as they once were. IBM, on the other hand, seems to be picking up a lot of new Linux, Unix, and Windows customers. Microsoft is number three in the market and
will continue be unless they decide to support different operating systems other than Windows.
And if you are a customer of one of these DBMS vendors, what does this stuff mean? In truth, it means very little. All three of these companies are very strong and very focused on the DBMS market.
You cannot go wrong choosing or using any of these products. My advice is to use them as leverage against one another. If you are an Oracle customer (for example) and you need to purchase
additional licenses, make sure your sales representative knows that you have spoken to IBM and Microsoft and are willing to convert. Tactics such as this can help to reduce your software license
Oracle Versus the State of California
The other prominent database news story of the past quarter focused on Oracle and a software deal they made with the State of California. Perhaps I should say, “tried to make” because the deal is
being questioned and state lawmakers are debating what should be done.
It all started when state officials signed a contract in excess of $100 million with Oracle. Then the state auditor, Elaine Howle determined that there was actually little need for the additional
Oracle software and that the savings claimed by implementing the software were dubious. It turns out that instead of saving $16 million the state might actually spend as much as $41 million more
than it would have without the new agreement. Whew! So who else is shocked that a software company overstated the savings potential of its software to close a deal? (insert sarcasm here)
This garnered a lot of media headlines for what it really seemed to be. To my eye, neither party looks very good. Between one party over selling and the other over buying there was a lot of
culpability to go around. In the end, it looks like the state of California will get some of its money back. And that is probably as much coverage as this sad tale deserves.
Oracle9i Release 2
In early June Oracle delivered Release 2 of Oracle9i. This new release provides several new and interesting features. First of all, Release 2 offers more native support for XML. Oracle9i now fully
absorbs the W3C XML data model into the Oracle database, and provides new standard access methods for navigating and querying XML – creating a native integrated XML database within the DBMS.
Oracle9i Release 2 also provides the next major release of the Oracle Internet File System (9iFS). This includes a new web interface to facilitate easier storage, collaboration, and publication of
content on a single Oracle 9iFS instance. New features in Oracle 9iFS include single file restore, workflow capabilities, and application plug ins to allow users to access Oracle 9iFS features
through Microsoft Office applications.
Oracle9i Database Release 2 also offers further support for Flashback Query, a feature introduced in the first release of Oracle9i. With Release 2 it is possible to generate flashback information
within a SQL statement, instead of just within a session. Additionally, the user can restore deleted rows, old values and the previous version of the table, as well as select a difference between
Other highlights include high availability enhancements, clustering improvements, Java 2 Enterprise Edition support, and support for Web technologies such as the Simple Object Access Protocol and
the Universal Description, Discovery and Integration directory of business to business services.
What Else is up at Oracle?
Well, the brain drain continues at Oracle. In mid April, Jeremy Burton, another high profile Oracle executive exited Oracle for a position at Veritas Software. Burton was Oracle’s senior vice
president of product marketing and he assumes the role of chief marketing officer for Veritas. At Veritas, Burton joins former Oracle executive Gary Bloom who left Oracle two years ago to become
CEO of Veritas.
Regular readers of this column will know that this is the fifth high level executive to leave Oracle in the past two years. The list of former Oracle executives leaving the company during that
timeframe: former COO Ray Lane (venture capital), Gary Bloom (Veritas), Peter Donnelly (Veritas), Michael Howard (Veritas), and now Jeremy Burton (Veritas). Well, upon further examination of this
list, perhaps this is not a “brain drain” at Oracle as much as it is a “brain suck” by Veritas? How many executives will depart Oracle for Veritas? Is Veritas becoming a mini Oracle?
In other Oracle news:
- The Walklett Group reported that the total cost of ownership of Oracle is twice as much as SQL Server over the course of five years. The study was conducted by analyzing the cost of running SQL
Server 2000 versus Oracle9i on a Unisys ES7000 box.
- In early June, Oracle announced general availability a partnership with Red Hat designed to promote the use of Oracle’s DBMS software on Intel based servers running Red Hat’s Linux
distribution. This included plans to optimize Oracle’s database and clustering software for Red Hat’s Linux Advanced Server operating system.
- In late June Oracle changed their certification requirements requiring new applicants to take a “hands on” class in addition to passing four tests. This significantly increases the cost of
becoming a certified Oracle DBA because the course costs about $2000.
Oracle’s Fourth Quarter
And finally, at least in terms of Oracle specific news, Oracle’s fiscal fourth quarter provided mixed results. Fourth quarter earning for the company dropped 23.3% from a year ago with net income
of $656 million on revenues of $2.8 billion. The previous year Oracle earned $855 million on revenues of $3.3 billion. For the year, Oracle’s net income was $2.2 billion (39 cents a share), which
is down 13.2% from last fiscal year (2001). Overall revenues for fiscal year 2002 were $9.7 billion, down 11.8% from last fiscal year.
What is so mixed about that, it sounds quite negative? On the positive side, Oracle met analyst expectations of 12 cents a share for the quarter. So even though revenue decreased, it did not
decrease any faster than expectations. This is a positive in today’s negative market. Additionally, Oracle reported its profit margin for the year to be 37%. That indeed represents an enviable
margin for Oracle Corp.
But to put things into another perspective, this was the first year in Oracle’s history that year over year revenue declined. With the heat on from IBM over the Gartner report and the year over
year revenue decline Oracle unleashed a press release in mid June detailing a slew of customers who purchased from Oracle in the fourth quarter. The list included some impressive customers
including America Online, Bank of Montreal, Bank One, Cox Communications, Nationwide Insurance, Pitney Bowes, and the United States Coast Guard. Hmmm… I wonder why the State of California was not
on that list?
There was little else as interesting this past quarter as the analyst’s DBMS numbers and Oracle’s on going shenanigans, but here are a few highlights:
In New Orleans at Microsoft TechEd, Microsoft outlined its vision for bringing .Net and web services to Yukon (the code name for the next version of SQL Server). This included plans to embed .Net
CLR (Common Language Runtime) in the SQL engine, as well as provide deeper and richer XML support. The.Net CLR integration will allow developers to write stored procedures in 23 languages in
addition to Transact SQL.
In other Microsoft news a flaw was reported in mid April that affected SQL Server 7.0 and 2000. The flaw involved a buffer overrun that enabled hackers to execute code on the machine where SQL
Server was installed. The flaw takes advantage of the way extended stored procedures validate input to the database. Microsoft has since issued a patch to correct the flaw (available at http://www.microsoft.com/technet/treeview/default.asp?url=/technet/security/bulletin/MS02-020.asp)
databases with no set system administrator passwords. When found, the worm sets the password to a random name and copies some files to the system. The worm basically gather password and database
configuration information. It is not destructive but it can cause an infected server to consume a lot of wasted resources and it can be messy to clean up.
And in DB2 news, IBM announced a new version DB2 OLAP Server this quarter. The new version, 8.1, offers hybrid analysis, which is comprised of relational and multidimensional OLAP. So users will
have real-time analytics across a broader range of applications.
Overall Market Demand
Finally, to wrap things up for this quarter, on June 17th Information Week reported that demand is not high for database software. The gist of the story is that a Morgan-Stanley Dean Witter report
indicates that most companies are not planning to buy more DBMS software, but instead plan to run the software they currently own. There are some other interesting factoids regarding DBMS loyalty
and needs that are well worth reading. The full story can be found at http://www.informationweek.com/story/IWK20020614S0026.
So that concludes another installment of the Database Report. The big three DBMS vendors continue to slug it out by interpreting analyst research, announcing (and pre-announcing) new features, and
selling (or over-selling) their wares to the data-starved masses. Tune in next quarter to stay abreast of what is going on in the wild, and woolly world of database management.