Oracle is working hard to become a one-stop-shop for customers' data integration, master
With more than 26 years of covering the enterprise applications market under his belt, Swanton should know.
SearchOracle.com caught up with Swanton recently to learn more about Oracle's goals and strategies with these increasingly important areas of technology. In this interview, Swanton sizes up Oracle's newly released
What do you make of Oracle's newly released Data Integration Suite?
Bill Swanton: Oracle acquired [Sunopsis Inc.] last year and its product -- [which is now called Oracle Data Integrator] -- had been used in Oracle environments. But what they're really doing with the Data Integration Suite is making [Data Integrator] fit better into the overall Oracle environment, and to some degree it is becoming a replacement for the Oracle Warehouse Builder.
Data integration is always a hot topic, but that's especially so these days. Why is that?
Swanton: Everybody needs to integrate all of the data sources that they have. Everyone is basically trying to leverage the information assets that they have more effectively, and more and more that means, "How do I get information into one place in one consistent way where I can work with it?" That is what Oracle is trying to provide here. They had some rudimentary capabilities before, and they expanded with a more general-purpose [extract, transform, load] tool. They basically are providing a way for people to use their information more effectively.
How would you characterize Oracle's overall data integration strategy?
Swanton: I think that Oracle, like everybody else, is trying to become much more of a one-stop shop. If you look at it from the Oracle Database and technology point of view, they're trying to be more like IBM, which BPEL Process Manager, which will let people take a string of applications and build one business process across them and thus do that process more effectively.
Another key area you cover is MDM. In your experience, what are some of the key pain points that organizations are facing today with regard to MDM?
Swanton: Users are still trying to get the MDM thing right. They're now realizing that they really have to have the right pieces there. The tools are just coming together to make that easier to do. Right up until now it has [required] quite a bit of manual effort to build something with workflow and everything else to do that. The tools that Oracle offers in R12 of the applications make it a bit easier for the customer with regard to product information. That's really the pain point -- there have not been a whole lot of good solutions out there that can be implemented readily and that work well with the applications, and that is just now changing.
Is it possible to achieve complete MDM within an organization?
Swanton: It's always going to be incremental. You're always learning something. Most companies have to start with a particular scope, which means a particular master data type and a particular scope of applications. For example, if you think of something like a product, product information is all over the company. Just in the ERP system from a supply chain and manufacturing point of view, a product might have 300 to 400 attributes that 15 to 20 people need to maintain. If you start getting into the marketing side of the world, if you start getting into the product development side of the world, you have as many or more attributes in each of those universes. So there is a huge amount of data, and people have to approach it incrementally. And it can be years and years of work because you have to get everyone to agree on how things are done and who is responsible for what, and then implement technical systems to do that. But it is something that people will eventually get to.
Do you have any general advice for people embarking on a data integration or MDM initiative?
Swanton: Like everything else, the hard part is the people. You've got to make sure that you have a governance process in place to agree on the rules. Technically implementing rules usually isn't anywhere near as hard. When we see people doing MDM programs, half of the effort is getting everyone to agree on what the rules are, 40% of the effort is cleaning up the data, and only 10% of the effort is actually implementing the technology.