In-Depth

Q&A: UDS—Informatica's Plug and Play Solution

Bottom line: IT can spend less time interacting with the business and more time dealing with strategic problems

• By Stephen Swoyer
• 07/14/2004

Two months ago, ETL pioneer Informatica Corp. announced Universal Data Services (UDS), a patchwork architecture that it positions as a kind of data integration Philosopher’s Stone: an extensible framework for data integration that facilitates the discovery, sourcing, cleansing, integration, and delivery services for any conceivable data source.

We spoke recently with Harriet Fryman, group director of product strategy with Informatica, about UDS and about Informatica’s ongoing effort to shrug off the dreaded “ETL specialist” moniker once and for all.

Fryman acknowledges that Informatica has had mixed results when it’s pursued strategies outside of its core ETL market—its analytics business didn’t pan out, but its PowerAnalyzer BI tool is doing well—but says that UDS is true to Informatica’s data integration roots.

You’ve quite forcefully resisted being branded as “just” an ETL vendor, although ETL remains your bread-and-butter specialty. With Universal Data Services, do you think you’ll be able to shrug off that label once and for all?

We are best known for PowerCenter [Informatica’s ETL offering], but what we’re describing with Universal Data Services is about providing a unified infrastructure, a unified way to interact with data so that IT can spend less time interacting with the business and more time dealing with strategic problems. So ETL is a part of it, but there’s a lot more there than just [ETL].

With UDS, it sounds like you’ll be attempting to grow beyond your traditional presence in data warehousing, or even business intelligence. Why do you think that IT organizations will be receptive to [a company like] Informatica in these extra-data warehousing-types of applications, where other integration specialists—the IBMs, Tibcos, and WebMethods—have long been dominant?

[UDS] is a product architecture, and our customers really see this architectural development as the way that they manage data in the future. It really gives them the opportunity to extend what they’re doing today and to innovate.

That’s not only in data warehousing (where people are very familiar that Informatica can provide value), but also in areas of ERP consolidation, system consolidation, and elsewhere. IDC identified the fact that CIOs have a lot to accomplish with their data, and we see it reflected in our own customers, because 30 percent of our customers are doing theses broad data-centric initiatives, and really there is a need and a desire to have this unified access to data. It’s all about getting that common interaction with data, and that’s where we’ve traditionally been strong [with ETL]

Your BI business appears to be pretty healthy, but some of your forays away from ETL haven’t been as successful. Last year, for example, you quietly exited the analytics business. Is there any reason why you feel that you’ll be able to expand into a new market and attract new customers with UDS?

First, our decision to exit the analytic [space]—OLAP vendors grew up to be business intelligence vendors, and they’re now trending towards providing high-end analytics through business performance management and corporate performance management. So we took our analytic applications business and went to partners with it, because we felt that partners were best equipped to address individual customer needs.

What that did was enable us to spend 100 percent of our focus on delivering these Universal Data Services. So we’re providing these Universal Data Services, and we’re looking to embrace and extend existing IT investments. So we provide the data discovery, data sourcing, [and] cleansing and integration services. We provide the services to cover this whole spectrum of data processing, and most importantly, the ability to audit that information.

But why Informatica? Why not SAP, with NetWeaver, or Siebel, WebMethods, Tibco, or others?

Because we’re providing a truly extensible architecture. Our data server [which provides performance, scalability, scheduling, and security, along with shared services] provides a single foundation into which the services of these products plug into, and we really feel that this is where we’re very different. We believe that service-oriented architecture is very important, and that’s why we have service-enabled PowerCenter [Informatica’s ETL tool]. We believe that that architecture is incomplete unless it’s founded upon a single robust data server that can do the processing of that.

You say you’ve designed UDS so that third-party vendors can plug into it, and in that way, you plan to flesh out the rest of your data integration stack, with third-party data quality technology and other components. Beyond what you’ve done with [data quality specialist] FirstLogic, is this something you expect will be popular among third-party vendors?

We’re seeing a very big trend happening where people are going to centers of excellence in order to take more of an architectural approach, so our shared services architecture really fits into what they’re trying to do with best practices.

We really see other partners embracing and extending our architecture, and this is where FirstLogic can embed themselves in this architecture as a shared sever. What we open up to [FirstLogic and other partners] is a way of accessing anything in [UDS]—we can provide the underpinning performance and scalability. This gives us a degree of flexibility that our competitors don’t have.

Because you can either develop in-house or partner for specific requirements?

Exactly. If we decided, for example, based on customer demand that we wanted to offer some EII or on-the-fly capability, we could partner or build our own, but immediately, it would sit on top of the data server and inherit all of the core values of that data server.

It’s easy to see how ETL tools like PowerCenter and PowerConnect fit into the UDS vision, but what kind of role do you envision for your BI offering [PowerAnalyzer]?

The way we talk about the ability to deliver data to applications, processes, and people, that’s where we see PowerAnalyzer as a very strategic product to us. Where PowerAnalyzer is very strong today, we’re adding to that capability for PowerAnalyzer to look at hierarchical data such as rows and columns.

PowerAnalyzer will report directly on information that other business intelligence products may not be able to access, for example, directly grab mainframe data [using Informatica PowerConnect]. That’s the way that PowerAnalyzer is fitting into our architecture and delivering a value that we think you have to have if you’re going to offer that complete interaction with data.