Managing big business data is big business itself
Originally published in Business in
Vancouver,
August 21, 2001, The high-tech office column
by ALAN ZISMAN
While most individuals needing to manage data
on their personal
computer either plug it into a premade personal information manager or
just dump it into a spreadsheet, managing corporate data is big
business.
In fact, with its self-named corporate database
software, Oracle
is right behind Microsoft in any list of the largest software
companies,
while company chair Larry Ellison is right behind Bill Gates
on lists of personal wealth (and reportedly deeply resents his
perennial
second-place standing).
While Oracle leads with some 33 per cent of this
lucrative market, IBM,
with its DB2 is not far behind. Microsoft's SQL Server trails with
about
15 per cent, followed by minor players Infomix (recently bought
by IBM), Sybase and others.
Market-leader Oracle is widely regarded as offering
the fastest product,
though each company manages to trot out a set of benchmark tests
revealing
their offering is the best-performing. Oracle, however, has also long
been
the most expensive. As a result, the company accompanied its newest
Oracle
9i with a new pricing structure, in order to maintain its share of this
$13-billion market.
Up until that version, Oracle had been charging
according to "power
unit" -- the faster the server that was housing its software, the more
a customer paid. Now, like IBM and Microsoft, Oracle will be charging
customers
per-processor. The same software, run on a quad-processor server will
cost
more than if it's run on a dual-processor server. Despite the changes,
Oracle remains significantly more expensive than DB2 or SQL Server,
though
the company suggests that its product's performance and features
provide
added value.
While it's clear that big databases will only get
bigger, the industry
is in for changes. Users want to access corporate data in more and more
ways: over the Web and from mobile devices such as handheld computers
and
even cell phones. And the data itself will be more dispersed as a
result
of software developments such as IBM's Web Services and Microsoft's
.Net,
existing on multiple computers across the network or even across the
Internet.
And, of course, both IBM and Microsoft would like to
have corporate
buyers decide that their respective database systems are the best fit
with
future Web developments. All three major vendors are busy integrating
XML
tools into their systems so that their databases will be able to flow
seamlessly
into the Web. (At the same time, they are hoping to head off challenges
from new products designed from the ground up around the new XML
standards.)
Of the three, Microsoft offers the lowest-cost and is
the easiest to
use. Its SQL Server includes strong data analysis features. On
technical
levels such as programmability, however, it is poorly regarded.
Not surprisingly, IBM gives the best support for
mainframes. The company
is also doing a good job of supporting Linux, the fastest-growing
server
operating system, and data warehousing. Its technology does well on
server
clusters. DB2, however, is more complex to set up and maintain than
Microsoft's
SQL Server.
Oracle offers the best performance on single servers,
but its new clustering
features aren't as developed as IBM's. It not only is the most
expensive
of the big three, but also the hardest to manage. Moreover, the
company's
reputation for customer support has been charitably described as
"inconsistent."
Managing their corporate data will remain the biggest
(and most expensive)
job for big business IT staff. But for the rest of us, it looks like
the
expanding integration of enterprise databases and the Web will make
more
information accessible in more ways than ever before. n
Two weeks ago, I mentioned that readers could
find a sample database
online listing all my writing. Of course, subscribers can access a
database
of all my BIV content, along with everything else in BIV
back-issues at www.biv.com.
Just
one more reason to subscribe.