IBM’s Informix purchase: Money for … what?

A billion here. A billion there. As the late U.S. Sen.
Everett Dirksen once said, after a while you’re spending real money.
IBM’s recent billion dollar purchase of Informix software looks like real
money — but it was spent on a dying company. Further, it was spent on
to support a software design that’s starting to look seriously out of
date. Go figure.You may have read that IBM has
just bought the rapidly shrinking remains of what used to be
Oracle’s chief competitor among major database software
publishers. A few months ago Informix began to divide itself in half, as it began
to devise a strategy to divest itself from its core database
business. On Tuesday, Big Blue acquired the assets of that database business,
leaving the rest of Informix to re-name itself Ascential Software.

The market responded by dropping IBM’s stock price a
couple of points. The analysts who commented weren’t negative but
they were cynical. The Gartner group suggested that the $1 billion dollar
purchase price represented little more than an
attempt to buy market share. Naturally Oracle dismissed the deal,
suggesting that IBM will likely inherit more of Informix’s recent problems than
much of 100,000 independent-minded database customers.

Most analysts think that IBM’s purchase was aimed at making sure that Oracle didn’t acquire those same customers, and was intended to turn the heat on the competition
just as Oracle’s sales had begun to flatten out.

A far more interesting question is why IBM’s leaders thought the purchase of a
dying Unix database firm with six separate and shrinking database
products would do anything for them. The relational database business is getting
rather interesting, and not necessarily in a good way, if you’re the
vendor of closed proprietary database software.

Traditional database systems have a problem: They’re old, and they
are beginning to look it. Increasingly, new applications are requiring a
new way of looking at data sets. The Web, and a variety of proposed
e-commerce and other data exchange applications require universal, incredibly
flexible and completely open database structures. It looks like the new format
for data exchange between these universal databases is going to be
XML, the shortened-up, Web-ready version of the Standardized General
Markup Language (SGML) that’s been used for years to express the layout and
content of electronic documents.

XML is basically a common format being used to
describe and wrap of content elements — only the old definition of
content is being rapidly broadened, to the point that the distinction
between documents and data is being effectively erased. The old definition of
wrapper is being broadened too, as descriptions have evolved into a form of
encoding that can guide how future programs interpret and manipulate the
content.

XML is being developed in an open environment, and, at least for the
moment, it has the same kind of messy, disorganized energy that Open
Source software developers have managed to get used to. New standards
committees are being spawned indiscriminately. There are lots of proposed and half
finished designs splattered all over the ‘Net.

The expected payoff for all this energy is a new set of highly
sophisticated applications such as Enterprise Information Portals (EIPs), which
are flexible frameworks for cooperative processing using application
elements such as data/document content management, warehousing
and archiving, content publishing, data mining and conditional
information sharing; a type of program that is really quite beyond the capabilities
of today’s standard database platforms.

The problem for many database developers is that, the more
application designers look at XML, the more traditional database systems look like
legacy platforms to be overcome. Instead of database platforms that
shoehorn XML into old-fashioned database tables there’s building demand for a
new database structures that make better use of native XML.

One such product is Open Source. And on the same day that IBM
announced it had spent a billion dollars on Informix, a fellow named Falko
Braeutigam announced the first stable release of the Open Source database called Ozone.

Ozone is a complete, object-oriented database
management system built in Java, and capable of running Java database
applications.

Ozone isn’t exactly new. An earlier version, Ozone O.61 has been
around for a while. This mature 1.0 may be stable, but it hardly has all the
needed features required to chase the likes of IBM or Oracle out of the
market just yet. In some ways it’s relatively primitive stuff, but it
has little of the legacy burdens borne by the commercial products. It’s
open enough to serve as a universal platform, and it is the first database
platform to treat XML properly.

At some level, most commercial vendors have tried to map XML back
to their proprietary back-end database systems to achieve much of
their functionality. The Ozone approach is is cleaner and simpler — and
a more logical starting platform for down line XML development. Others
are developing additional content management layers among other
elements.

Ozone includes a fully W3C compliant Document Object Model
(DOM) implementation. DOM is a platform- and language-neutral interface that
allows various programs or scripts to dynamically access the content structure
and style XML-wrapped content through an open DOM interface rather than
a proprietary, and product-specific APIs. Product developers can use
the DOM interface to cooperate with each other, increasing type of
interoperability required to realize required to develop a variety of
cooperative processing and complex B2B applications.

Ozone may not be ready to compete with Oracle any more than
Informix apparently was. But it is a logical Open Source platform on which to
build the future “universal database” required by next-generation XML
applications. And it cost a whole lot less than a billion dollars.

NewsForge editors read and respond to comments posted on our discussion page.