This year is the year IBM has chosen to get more
intelligent about business intelligence.
While the company has delivered BI-related applications and tools
for some time, IBM is now making a commitment across all of its
major software groups to deliver core products with a healthy dose
of BI baked in.
One of the major goals of this multi-division effort is seamless
integration of the massive stores of historical data that corporate
users typically have with real-time analytics. Company officials
see this as mandatory if corporate users are going to establish a
true on-demand environment successfully.
Janet Perna is the IBM executive helping piece this strategy
together. As IBM's general manager of data management solutions,
Perna has been in charge of IBM's multibillion-dollar DB2 business
since the mid-1990s and was a key player in IBM's $1bn acquisition
of Informix in 2001 which doubled IBM's distributed database
business overnight.
More recently Perna has been helping shape and deliver the
company's next generation of database and content management
products that will comprise the infrastructure for creating an
on-demand environment. Perna discussed some of IBM's latest
directions in this market.
IBM has been in the BI market for quite a while, but the
company's On Demand initiative appears to have given it renewed
purpose. What is the strategic intent of BI for IBM?
Perna: If you look at what companies are doing as they
integrate their systems horizontally, they want to optimise their
overall business processes so they can respond to changes quickly.
It allows them to quickly maintain their existing business
processes based on both internal and external factors. In order to
do that you need to have insight as to what is happening, and that
really everything to do with intelligence.
Can you give an example of this?
Perna: Look at Laughing Cow Cheese, which is mentioned in The
South Beach Diet, but not by its specific brand name. (Laughing
Cow) had no idea that the brand was being referenced in the book;
that was something happening external to their business. Imagine if
they had known that.
You start thinking about technologies like Web Fountain that
allow you to go out into the community of external sources and get
insight into what is happening out there that might affect your
brand and business, either negatively or positively.
So when you think about this whole notion of On Demand and what
companies are trying to do, BI gives you business insight through
information on all types On Demand data.
Web Fountain is a technology we have now in (IBM) Research and
one of the things it does is crawl many sources on the web to
gather data that might be relevant to your business. One of
the things Laughing Cow would have been able to realise was their
brand was mentioned in a best-selling book, and so if a lot of
people were asking for it at the grocery store, they could
anticipate that added demand and been ready for it.
What is IBM's strategy for merging the massive amounts
of historical data a large company has with real-time analytics? Is
it just an extension of your existing integration strategies or
something else?
Perna: Analytics are just going to be part of applications like
applications for ERP, supply chain, and CRM. We are beginning to
see that evolve today, as we see with business process integration.
What we will see is events triggered by data. There will be changes
to a state in the database that are going to trigger some kind of
behaviour.
It is at that point where you integrate the analytics. But part
of the requirement for doing that is the database be smart enough
to be able to warn the analytic functions that something is coming
in.
For instance, when a web-based transaction is coming into the
database, you want to be able to score this person I am transacting
with in real time using data mining technology. So if this person
is a great candidate for a life insurance policy or trying to get a
seat on flight, the transaction you are conducting will invoke
triggers to call up real-time scoring services, which gives you
more information on the profile of this person you are dealing
with.
Think of it as a kind of conversation, or taking CRM back 100
years ago. When someone walked into a merchant's store 100 years
ago, they knew a lot more about that person than merchants do
today. Back then they would know that person's birthday, family
background, etc. Well today we try to have that sort of
relationship with customers collecting that sort of information
through many, many transactions.
This real-time scoring software is Masala?
Perna: Yes. In essence what we are trying to do is interact in
real time as if you and I were talking. Well, not so much talking
in this instance, but interacting with a system and having that
system and data be intelligent enough to figure out what is the
next thing you want to do.
How complex a task will it be to get people to upgrade
their infrastructure to marry all their historical and real-time
analytics?
Perna: There is not a big rip and replace strategy needed here. Our
approach is really about leveraging infrastructure you already
have, wherever it exists. If you look at all the data sources out
there, it is in file systems, content sources, a range of web-based
information.
And so when you look at what we are doing with the Information
Network Integrator, our integration platform for information, we
have been extending its reach. Initially, it would go out to just
relational data sources. Then, we added e-mail, web services, and
XML data sources, and now we are adding text. We have this search
paradigm with Masala to do free-form search. So what we are doing
is using that layer to search and integrate and bring in
information from all these different sources.
How closely do you have to work with Microsoft to make
all this work effectively with things on the desktop? And what sort
of standards do you both have to agree on to make that
happen?
Perna: Most of the standards are in place and would include SQL,
XML, and XQuery. So when you look at what we have here, we have the
necessary integration layer, the programming interface layer is
SQL, and with Masala you have enterprise-level search. These are
all different metaphors for accessing information.
So you see the DB2 Information Integrator as the key
layer in terms of expediting this strategy?
Perna: Exactly right.
So what is the next step for carrying DB2 Information
Integrator forward?
Perna: We put Masala into beta over three months ago. This will
give us an enterprise search paradigm. Think about a portal app
where you are not dealing with an app but a person, and the
metaphor is free-form search.
The other thing we have already integrated is text mining. So if
you want to know information about which appliance is calling into
your call center, today what people have are records, problem
records… .
What will portal technologies play in this
context?
Perna: To me portals are one of the ways to get it all out to
the masses. So the (Lotus) Workplace environment is fast becoming a
personalised portal environment. Portals will be one of the places
where you can deliver this information.
There is this term "business performance management," and what
that is really about is, we have this end-to-end business process
that starts with CRM and goes to supply chain management systems
and the optimisation of that whole process. So it is taking this
end-to-end view and this business model and applying analytics to
that end-to-end process that shows where the bottlenecks are in
that process that slow the flow.
Once that is done, companies can figure out where and how to
change the process. After a while, the process can change itself to
better adapt through real autonomic technologies. So this layer
gets more and more automated, which means you can gather
information at every step of this business process and the
information flows from one business process to the next, you can
analyse that along the way.
Microsoft has to play some role in all of this to push
data down to the masses, but they are not a big pusher of BI
technology at this point.
Perna: Right. Microsoft has the dominant (desktop) applications
suite. They use Excel as their BI tool, which is fine. And so the
connection we want is Excel sitting out here and being able to
access this infrastructure. We have the ability today to gain
(desktop) access through Office.
What other changes have you made internally to help
accommodate this renewed BI strategy?
Perna: One of the biggest changes we made this year was the
restructuring of our sales force to better focus on industries but
also to better focus on solutions. Now we have a BI speciality team
that spans across IBM. And because we have an open standards-based
strategy (with BI), if users want a portal strategy for Linux on
the pSeries, we can do that.
Ed Scannell and Tom Sullivan write for
InfoWorld