Well, another quarter has gone by and many skirmishes have been fought in the database wars. As usual, Oracle is making the most noise and has the most news. Also, as usual, the news is less
technology-focused and more about marketing, sales, and revenue. So, we’ll look closely at what has been going on at Oracle, but we’ll delve into the some of the database-related news coming
from, Microsoft, IBM, Sybase, and others too.
You can’t pick up a newspaper or visit a web site these days without hearing about the slowdown of the US economy and the oncoming bear market. But what about the DBMS marketplace, is it as bad as
the market in general? And do the stock prices of the major DBMS vendors truly indicate the end of the database wars?
Well, let’s break from conventional wisdom and take a macro look at DBMS sales. By most reasonable measures DBMS sales continue to increase. Oh, not by the huge percentages by which they used to
grow. But the growth is nothing to sneeze at! According to AMR Research, recent year to year comparison of quarterly results showed 50% gains for Microsoft SQL Server, 20% gains for NCR Teradata,
and even 10% gains for Sybase Adaptive Server. Even Oracle’s results for its fiscal second quarter were not really that bad with regard to Oracle8i sales: a 19% increase over the same quarter the
previous year. And IBM reported year over year fourth quarter growth of 17% for DB2 for OS/390.
In another good sign for the database industry IBM reported in March 2001 that it exceeded 10,000 licenses for its DB2 for OS/390 DBMS. According to analysts at Dataquest IBM owns 96% of the
mainframe DBMS market. Further, IBM notes that 80% of the Fortune 1000 use DB2 for e-business.
So there is a lot of good news for database users. The market is healthy and continues to grow. But, as with all things, the real news is in the details. Let’s look at some details, starting with
Oracle’s Third Quarter
Oracle had its share of difficulties during the first quarter of 2001 (during which Oracle reported its fiscal third quarter results). In early March Oracle pre-announced that third quarter sales
would be lower than anticipated because “a substantial number” of its U.S. customers slowed down their IT spending. Oracle’s message was quite clear, database sales would suffer because of the
uncertain U.S. economic climate.
On March 15, 2001, Oracle posted its third quarter results. The results were in line with Oracle’s revised estimates: third quarter income for the company increased 16% to $583 million, or 10
cents a share. For the same quarter in 2000, results were $503 million, or 8 cents a share. Overall revenue for the quarter was $2.7 billion versus $2.4 billion for the same quarter last year. All
in all, to me this seems like healthy, if not spectacular growth.
But we have come to expect spectacular performance technology companies, and those expectations were priced into the stock of companies like Oracle. Well, until recently that is. Toward the end of
the first quarter Oracle stock (ORCL: NASDAQ) was trading at just under $16 per share – more than 30 points lower than its 52-week high.
Oracle reported 6% quarterly license growth rate for database sales. This growth rate was higher than the slightly positive or slightly negative database growth rate suggested in early March. But
it was much lower than Oracle’s original 20% growth prediction. Even sales of Oracle’s applications rose slower than expected – 25% instead of earlier predictions of 50% to 75% growth.
Oracle blamed the slower growth on economic issues and claimed that it really did not know sooner that growth would be so severely impacted because many of their deals close late in the quarter.
There is a lot of truth in both of those statements, the latter more so than the former though. Most businesses that buy enterprise software try to postpone purchasing decisions to the end of the
quarter to negotiate better terms and conditions on their licenses. The thinking goes something like this: “I bet that vendor will be more likely to reduce their prices if they are in danger of
missing their quarterly sales targets. We need their software, but let’s just try to wait them out.”
So why did Oracle’s sales slow? Was it really all just attributable to the economy? Well… on March 16, 2001 Computer World reported that some customers blamed Oracle’s pricing policies. Some
Oracle customers believe that Oracle prices its software too high and is difficult to negotiate with. The Computer World article cited Teri Palanca, an analyst at Giga Information Group, backing up
the customer claims saying “we believe many customers are at a minimum elongating their sales cycles, if not completely deciding against Oracle and moving to competitors, solely based on cost.”
Furthermore, Eric Upin, an analyst with investment banking firm Robertson Stephens, cites further reasons for Oracle’s shortfall including “sales execution; the potential impact of management
departures…; overly aggressive original guidance…; a major product rollout in the 11i applications suite, and of course, the slower economy.” I would add a few things to this analysis.
The packaged COTS (Commercial Off-The-Shelf) software vendors of ERP and CRM solutions, notably SAP, Peoplesoft, i2 Technologies, and Siebel no longer promote Oracle as their primary platform.
Instead, most have indicated their preference for IBM’s DB2. Since Oracle competes with these vendors with their own suite of packaged applications, the COTS vendors decided to back a different
DBMS. The impact of Oracle no longer being the preferred database for these applications cannot be helping Oracle’s revenues.
And though I understand Oracle’s reluctance to admit it, I’m sure pricing is an issue for most customers. Oracle prices their DBMS product very aggressively, which is to be expected because they
have been the market leader for quite some time. But competition has increased. The battle of the “old days” was easier for Oracle when they were primarily fending off smaller companies like
Informix and Sybase. Today Oracle’s primary competition comes from 400-pound software gorillas like IBM and Microsoft.
Oracle’s marketing may also have contributed to the less than stellar DBMS revenue. By promoting Oracle9i well in advance of its scheduled availability potential customers are likely to delay
purchasing the current version, Oracle8i, until the new and improved version is available. Of course, such pre-marketing and promotion is nothing new for the software industry.
Additionally, Oracle’s largest customer base runs their DBMS software on Unix platforms. There is heightened competition for this business. It comes from IBM with DB2 running on OS/390 and the
combination of NCR and Teradata at the high end; and on the low end, the inexpensive value proposition of Microsoft SQL Server running on Windows NT/2000.
And, yes, just like Larry says, the slowing economy is a factor here, too. If you’d like, you can review Oracle’s actual earnings report at http://www.oracle.com/corporate/press/index.html?624479.html
In the Wake of the News
As a reaction to the lackluster earnings announcement, in late March 2000 Oracle reported its intentions to cut the size of its workforce. In a press release Oracle stated that “Based on current
business conditions, at this time the company expects to reduce our worldwide workforce by approximately 1%-2% through normal attrition and regular business performance assessments, in line with
our ongoing global e-business process improvements.” With about 43,000 employees worldwide, the cut could impact about 860 jobs.
The Exodus Continues
Well, Oracle got a little help toward achieving their desired attrition. Veritas, who last quarter lured Gary Bloom away from Oracle to become CEO, has snagged two more Oracle executives. So the
brain drain we discussed last quarter (Ray Lane, Gary Bloom) continues this quarter with Peter Donnelly and Michael Howard, both of whom were vice presidents at Oracle.
Peter Donnelly was vice president of finance at Oracle and was named vice president for business process reengineering at Veritas. Michael Howard was vice president of B2B integration and the data
warehouse division at Oracle and was named vice president of the Internet division at Veritas. You have to figure that Oracle really didn’t want these two guys to leave as part of their 2%
Okay, But What About the Technology?
Well, we’ve spent a good amount of time analyzing Oracle’s business, but have not yet discussed Oracle’s new technology initiatives. In the first quarter of 2001 Oracle unveiled its new Data
The Data Guard technology enables an organization to set up a standby database at a remote location. Every transaction against the primary database is replicated to the remote database. If an error
or failure occurs, operations quickly can be switched to the standby database using an automated process. Additionally, the standby system can take over while DBA maintenance work is performed on
the primary database system. Data Guard delivers higher availability, which is much needed in this e-business, web-enabled world. Data Guard technology will be built into Oralce9i and is available
as a free add-on option for Oracle 8i.
And despite its financial troubles, Oracle has not changed its plans to release the new Oracle9i DBMS software by the end of June 2001.
On the publicity front, Oracle got some good coverage of its current DBMS release, Oracle8i Release 3, in eWeek. An eWeek Lab Review of Oracle’s database stated that only IBM’s DB2 can compete
with Oracle8i’s breadth of database functionality. The review further touted Oracle8i’s strong integration with the Internet claiming Oracle8i’s IFS (Internet File System) to be “brilliant and
original” and having the ability to make content and file management much easier. But eWeek questioned Oracle’s performance. As most Oracle users know, Oracle’s license prohibits publication of
independent performance studies, so we’ll never know.
Nonetheless, the overall eWeek review was quite positive with Oracle8i Release 3 receiving an “A” grade for capability, and “B” grades for usability, performance, interoperability, and
manageability. Check out the review for yourself at http://www.zdnet.com/eweek/stories/general/0,11011,2666959,00.html
That DBMS from Redmond
Oracle was not the only company making news this quarter. Microsoft and its SQL Server DBMS both were very active in the news, too. Microsoft settled a long standing dispute with Sun Microsystems
over Microsoft’s use of Java, many Microsoft web sites experienced long outages due to technical glitches and denial-of-service attacks, and Microsoft’s appeal of the anti-trust decision that
would split it in two began to heat up. But none of these things really impacts the DBMS world too much.
On a more significant note, Microsoft acquired Great Plains Software. Great Plains is a vendor of COTS software applications. The acquisition should help Microsoft to compete more broadly against
Oracle in the enterprise software space – Oracle offers both DBMS and application software. Now, so too does Microsoft. Great Plains customer base consists of mostly small- and medium-sized
businesses that are likely to reject the premium pricing of Oracle’s software. If Microsoft can gain some synergy between the Great Plains applications and SQL Server they can offer powerful
solutions for small- and medium-sized companies – a market that Oracle would like to serve, too.
.Net? Not Yet!
In other database-related technology news from Microsoft, the company is touting .Net (pronounced dot net), a new scheme for providing software-as-a-service. Microsoft .Net is basically a set of
software technologies to help build, deploy, operate, and integrate services over the Internet. Using .Net, applications become Web services that can be delivered over the Internet and combined
like building blocks to create customized solutions.
According to Microsoft’s plans, .Net will use messaging and XML to allow any application or device to share any piece of information and act on it, often without user intervention. But as of
today, .Net is not deliverable. It is Microsoft’s vague future vision for software. If .Net is successful it will enable you to connect all your data, devices, and applications together. An
auspicious goal, no?
From a database perspective, .Net introduces ADO.NET for data access as a replacement for ADO. ADO.NET uses XML for data transport, so ADO.NET can treat anything that emits XML as a data source.
Using XML may be a double-edged sword for Microsoft though. It is true that XML is a burgeoning standard and using it may enable ADO.NET to access more diverse data. However, XML adds a lot of
“stuff” to the data that is being transferred, which could impact performance for high volume data transfers.
All in all, though, .Net is a good idea and Microsoft has put together some interesting plans. But isn’t the goal of .Net trying to achieve what Java and J2EE were supposed to accomplish? Well,
yes and no. Both J2EE and .Net are development frameworks for assembling components in web applications. The difference between the two is where the portability is delivered. Sun’s J2EE enables
portable deployment of applications to multiple platforms (Linux, Windows, Solaris, OS/390, etc.) as long as you develop the application in Java. Microsoft’s .Net enables portable development of
applications in any language (Visual Basic, C#, C++, potentially Java and others) as long as you deploy the application on a Windows platform.
So, if you are content with deployment on Microsoft platforms, .Net may be just the thing you are looking for. On the other hand, if you need to deploy on non-Microsoft platforms but can get by
coding everything in Java, the J2EE should fit the bill. In the real world, though, wouldn’t it be nice to have both types of portability?
Suffice it to say, skepticism is a virtue when dealing with frameworks – especially frameworks that have yet to be fully delivered.
Accolades for SQL Server 2000
Microsoft was a recipient of some good news for its DBMS product, too. Datamation named Microsoft SQL Server 2000 the winner of its Product of the Year for 2000 in the Data Warehousing &
Business Intelligence category. Users gave Microsoft SQL Server 2000 high marks for security, responsiveness, and scalability. But in my opinion this award is a little weird. SQL Server 2000 was
competing against Accrue Insight 5, Brio.Portal, Actuate’s e.Reporting Suite, Metagenix MetaRecon 2.4, Sagent Portal, and SAP Business Information Warehouse, 2.0… not a DBMS in the bunch. I can
see where a DBMS with BI features might easily beat a BI tool without DBMS features, can’t you? At any rate, you can read the entire story at http://itmanagement.earthweb.com/datbus/article/0,,11969_600151,00.html
IBM: The Mainframe Catches Up
The big news out of IBM this quarter is the general release of DB2 Version 7 for OS/390, which happened on March 30, 2001. V7 of DB2 UDB has been available for Windows 2000/NT and Unix/Linux
platforms for at least half a year already. So is this news really positive for IBM? In fact, it might be reasonable to surmise that IBM is behind the curve in supporting DB2 on the mainframe. With
10,000 DB2 licensees running on their mainframes you would think that IBM would be a little more up-to-date with their DB2 version releases for that platform. But with IBM is intent on competing
with Oracle for Unix and Windows business, perhaps IBM is taking their mainframe business for granted.
Let’s take a quick look at the features IBM delivered with their mainframe edition of DB2 V7. First of all, V7 delivers XML support via a new data type extender for XML documents. The XML Extender
enables XML documents to be integrated with DB2 databases. It enables the storing and searching of entire XML documents using SQL, as well as the ability to break apart XML documents into columns
of database tables. Using the XML Extender, XML documents can be combined with traditional data stored in relational tables.
Other enhancements bolster application development and programming. DB2 V7 enhances stored procedure support with a point-and-click stored procedure builder, as well as providing support for Java
and SQL stored procedures. Other application enhancements include support for scrollable cursors and SQL row expressions, the ability to limit the number of rows returned by a query, the ability to
set external save points within a transaction, and numerous other small upgrades.
DB2 V7 also provides some much need enhancements to DB2 data management capabilities including support for identity columns, better support for temporary tables, and Unicode support. IBM has also
improved DB2’s built-in data warehousing and business intelligence capabilities
To summarize, V7 of DB2 for OS/390 is not as big a release as V6 was in terms of new features and functionality, but there are many new and useful features. IBM has done a lot of work to make the
various DB2 editions they support across various platforms more consistent in terms of features and functionality. For mainframe folks, though, the big concern may be that the OS/390 version of DB2
was the last to be released. Of course, with 10,000 licenses IBM owns the mainframe DBMS market. And DB2 for OS/390 is a stellar performer with enviable RAS qualities. Let’s hope IBM does not lose
sight of this as it attempts to combat market-leader Oracle on its Unix and Windows turf.
Accolades for DB2
A recent report from D.H. Brown and Associates touts the lower total cost of ownership achieved by companies running IBM’s DB2 as compared to Oracle. The report goes on to state that DBA tasks can
be accomplished more efficiently with DB2 than with Oracle8i – from 6% to 20% more efficient depending on the type of databases being supported.
“DB2’s pricing structure has a scalability advantage” states the D.H. Brown report. “…a DB2 UDB system often costs as little as one third the price of an equivalent Oracle8i system.” The
report seems to be fair and unbiased, but I have to note, every page of the report is stamped “Prepared for IBM.” So, I have to wonder, did IBM finance the production of the report? And if the
report was put together for IBM were there any negative aspects of DB2 as compared to Oracle that were not included in the report? Even so, this report is a powerful weapon for IBM in promoting DB2
Sybase: They’re Not Quite Dead Yet
Although Sybase has experienced some hard times during the past few years, the first quarter of 2001 was undoubtedly quite good for Sybase.
On February 20, 2001, Sybase announced that it will acquire New Era of Networks (NASDAQ:NEON) in a stock-for-stock trade valued at approximately $373 million. NEON is not the same company as Neon
Systems (NASDAQ:NESY), the systems management company in Sugar Land, Texas. NEON, an application integration company, saw its stock price drop from a 52 week high of $95.25 to about $6.00.
The move enables Sybase to bolster its arsenal of development tools. It will help Sybase to grow out of its traditional position as a DBMS vendor. The acquisition should add to Sybase’s bottom
line and enable cross selling opportunities – selling NEON’s products to Sybase’s DBMS customers, and vice versa. The combined company will have approximately 5,500 employees and more than 43,000
Sybase Declares War
In mid-March 2001 Sybase CEO John Chen slapped Oracle with a verbal attack. Chen intimated that Oracle has faltered in the database market because they have been paying too much attention to the
application market. In an interview with silicon.com Chen also criticized Oracle’s pricing policies claiming “Oracle (has) a complex pricing policy and at the end of the day they’re running that
policy to get more money. Companies like Samsung are sending a signal to Oracle that they’re not the only company to play ball with.” Samsung is a Sybase customer.
This is an interesting public relations ploy, but Sybase is so far behind in the DBMS market that it has little chance of catching up to the market leaders again. It is not just Oracle that Sybase
has to worry about, but also industry Goliaths IBM and Microsoft, as well as other smaller DBMS vendors like Informix and NCR.
And What About Sybase’s Technology?
In the first quarter of 2001 Sybase released a beta of version 12.5 of their stalwart DBMS, Adaptive Server Enterprise (ASE). Billed by Sybase as the e-Business Database, ASE delivers some nice new
features for enhancing productivity, management and security.
ASE 12.5 supports new Internet security and directory services, enabling users to rapidly build and deploy e-business applications. ASE 12.5 also delivers better security with support for both
Secure Sockets Layer link protection and row-level security. Additional enhancements help developers to quickly integrate XML-based applications with the ASE 12.5 databases and enable Java
developers to quickly create business rules and execute Enterprise Java Bean components in the database. ASE 12.5 also supports the Unicode encoding scheme.
News from Other Fronts
Even though the major DBMS players dominate the news, there are other significant DBMS products and vendors – some of whom made news during the first quarter of 2001. One of them, surprisingly
enough, is Borland InterBase. I say “surprisingly” because until just recently Borland did not even exist and InterBase was destined to become “open source” software.
Some of you may recall Borland from the early days of PC software. Borland created the terminate-and-stay-resident software category for DOS programs with Sidekick, its personal assistant
application. Borland also made a reputation as a programming language and compiler company with Turbo Pascal, Turbo C, and various other Turbo compilers, and later with Delphi and JBuilder.
Unfortunately, they were overcome with delusions of grandeur and tried to compete with Microsoft – which was their big undoing. They acquired a personal database, Paradox (since sold off to Corel),
a spreadsheet, Quattro (also sold off), and a word processor that did not last long. But more to the point, Borland also acquired an enterprise DBMS, namely InterBase.
But the losses mounted and Borland’s strategies changed. InterBase, at first a technology leader, languished under Borland’s care. Then Borland tried to change its name to Inprise, then wised up
and changed it back. In January 2000, while they were still calling themselves Inprise they toyed with the idea of going the open source route with Interbase. But Inprise continued to sell and
support the main version of Interbase through its normal distribution channels.
Then they changed their mind again. Now, calling themselves Borland again, InterBase 6.0 has been released and certified for Linux, Solaris, and Windows. Features include ANSI SQL, ODBC, and JDBC
interfaces, close integration with Borland tools, better data type support, enhanced utilities, and claims of lower maintenance requirements. So perhaps InterBase is back on track – good news for
its dwindling user base, but unlikely to win it many new converts.
Well, that brings the Database Report to a close for another quarter. But I’m sure things will remain hot and interesting during the next quarter. The major players are battling one another harder
and faster than ever before. And it is not just IBM, Oracle, and Microsoft making the news. So be sure to check in again next quarter to keep up-to-date on the ongoing saga of the database wars.
Until then, happy databasing.