Imagine trying to pull a single file from several repositories filled with
millions.
To make it even more daunting, let’s say those repositories were constructed
from different pieces of software that can’t necessarily communicate with
each other.
IBM is fusing several pieces of software from several recent acquisitions to
tackle this challenge, led by Steve Mills, senior vice president and group
executive for IBM’s software division.
Mills and several company executives convened in New York last week to unveil
the fruits of their labor and their next foray: a $1 billion investment in
technology and services to help customers get more out of their
information assets.
The executive added color to IBM’s information management strategy during a
sit-down with internetnews.com after the announcement.
Q: Will the $1 billion investment include more acquisitions?
We always go through these “make-buy” kinds of valuations, and we go after
companies whose capabilities we think can add to the portfolio. We have done
base technology investments on feature, function and performance, the
classical things you have to do to have a complete portfolio.
For the last
five years we’ve been accumulating or building the technology to deal with
all of the data and data types. How do you move the data, how do you
reconcile it and translate it into a common form? We’ve made a number of
different investments that have to do with data discovery and manipulation.
Acquisitions like DWL, Trigo and SRD focused on the domain-specific areas.
We’ve put together a pretty rich portfolio.
The next period of time will include some amount of base technology. But
more of it’s going to be focused on applied solutions.
So as we engage with
law enforcement agencies, states and their programs, and a variety of
others, a lot of the money is going to be going into how we take the pieces
we have and put together hard-knit, but repeatable, solutions and add the
right front-end function so administrators and users can use the technology.
This is where the money will go for next three-year-plus period. Within
that, we constantly have our eyes open for some company out there that can
enhance our portfolio.
Q: You told us about a New York Police Department composite scenario where
IBM’s software was used to narrow down suspects in a shooting based on
characteristics such as a tattoo and a nickname. Did the NYPD use the new
WebSphere Information Server and the WebSphere Content Discovery Server for
this case?
Yes. Their cumulative record base approaches 100 million records, so it’s a
pretty big database. The idea was to rapidly put together the right
information, do the compares and deliver it to a squad car on the spot.
Speed was a critical issue. In some cases, they’re literally thinking the
perpetrator is in the neighborhood. They want to get this guy right away.
It’s a classic information integration challenge, coupled with a basic set
of comparison algorithms.
Q: IBM mentioned a $69 billion market opportunity for information management
software through 2009. Do you see a lot of business for this coming from the
law enforcement side?
We think so. We think that at the state, local, federal and international
levels, they’re all dealing with some kind of fraud and abuse.
The size of
that market is anybody’s guess. You could end up with a trillion dollar loss
pretty quickly if you add up fraud crimes in all of those programs together.
We think a lot of money is going to go into government appropriation for
those kinds of projects because the demonstrated payback is so significant
that they’re all going to want to jump on the bandwagon.
Q: Since customers are dealing with so many disparate pieces of software,
what are the challenges of tying information management to service-oriented
architecture?
The notion of service orientation touches people, the business process, and
the information. So if I can’t do semantic reconciliation of data, then I
probably can’t integrate the process because the process has the data
associated with it.
There is the information itself that has to pass. What
form does it have to be in as it gets passed down the line from the process?
When nobody has a green field, they have multiple applications and expect
them to be in different forms.
So customers need to reuse the data in that
integration strain in order to deliver a level of service orientation.
Sophisticated customers who have been into this for awhile will tell you
that, for the governance and internal costs of how to get the shared service
approaches, the data problem becomes the biggest single problem they deal
with. The lack of common metadata structures drives them crazy.
Q: IBM seems to be corralling this information integration market more than
rivals such as Microsoft, Oracle and SAP. Is that a fair assessment?
There’s not much evidence that this is where they’re going. Oracle has
obviously become obsessed with the application market space so they’re in
competition with SAP. They’ve got their hands full with companies they’ve
acquired.
Oracle’s always had a view that they want to put all of your data
in the Oracle database. They’re not investing in federation technology to
deal with lots of data and lots of sources. So, they’re not really prepared
to compete in this market.
SAP talks about their application integration
hub, the notion of wiring everything into SAP. This would tend not to be
their area of interest anyway. They’re in the master data circle to some
extent, but a lot of this stuff, they’re not too focused on.
Microsoft tends
to be into more classical business and statistical analytics, as opposed to
complex data-mining scenarios that deal with different platforms across
engineering and scientific domains.
We can do the math against large amounts of data from multiple disparate
sources very quickly. We’ve been accumulating the pieces that let us go
after opportunities in ways that major software companies haven’t addressed
at all.
This is an opportunity that IBM is very well positioned to take
advantage of.