IBM to Add Native XML for DB2

IBM is developing a new XML-based storage manager for its
database software, highlighting what one analyst said is an industry shift
toward recognizing the database as a significant part of the infrastructure
again.

IBM will likely make the native XML storage manager available
in the next version of its DB2 Universal Database for 2005, said Jeff Jones,
director of strategy for data management solutions in IBM’s Software Group.

“We are recognizing that XML is a force that is rapidly growing for
organizing and representing any kind of information,” Jones said. “It’s not
just for document interchange or Web site painting. That force means there
is an increasing amount of XML in our shops, and the database needs to be better
structured to handle XML.”

The problem, Jones said, is that modern databases are better suited to
handle relational data in tables, rows and columns. Applying the relational
model to XML code means the XML has to be “shredded” or manipulated to work
in relational columns. This means XML loses much of its effectiveness.

While pure XML databases exist to handle the language, they are not fast
enough, nor do they scale large enough to handle enterprise workloads, Jones said.
With the new storage manager, IBM is making it possible for DB2 to handle
XML code as if the database were constructed out of XML itself.

The tool is not a relational table manager — it’s a second storage manager
that will function as both a relational table manager and an XML manager.
Jones and his team have solicited a select group of beta testers to run the
new storage manager as an add-on to DB2. No timetable has been set for the
inclusion of the native XML enhancements.

Jones said IBM is also providing developers more choice in terms of
programming language as they may choose to use traditional SQL to write to
DB2 or XQuery , which he calls “the most flexible means to
address XML with questions.”

The importance of such XML-enabling capabilities can’t be understated.
Many developers are using XML to construct Web services applications and
service-oriented architectures , so it makes sense to have
databases that can store the XML information from these applications without
shredding documents and putting them into columns.

Burton Group’s Peter O’Kelly said IBM, Oracle and Microsoft are leading a
resurgent database market, which languished in popularity after taking a
back seat to applications servers, which were red hot during the last few years.
With IBM’s new native XML enhancements, database software will likely regain
some of its lost luster, O’Kelly said.

Comparing their level of complexity to operating systems, databases
do such SOA-related tasks as message brokering, native Web services
support and event management, O’Kelly said. They also ensure corporations remain compliant
with record retention policies.

“When people go back and check their assumptions, they will see an expansion
of what you can do with DBMS because of XML and standards like XML Schema
and XQuery,” O’Kelly said. “You want to be able to take those data-centric
things in XML and put them into a database without a loss of fidelity, and
this is one area where IBM is going further than Oracle and Microsoft.”

Sprucing up XML capabilities in DB2 is just part of the company’s plan to
shoot past rivals Oracle and Microsoft in the competitive database market.
In the past two years, all three companies have added automated management
features and better search capabilities for their products.

Aside from enhancing DB2 with XML, IBM also has grand plans to boost the
product’s search capabilities. The last release, DB2 8.2 (code-named
Stinger) featured OmniFind, the search component of DB2 Information
Integrator.

OmniFind, a tangible result of IBM’s search research and development
efforts, helps applications find pieces of answers to questions from
multiple vendor databases and content repositories.

“It gets the right info
closer to the people who need it,” Jones said.