Q&A: Data Integration Platform, Not the ETL Tool, Is Key

Informatica isn't worried that Microsoft and Oracle are enhancing the ETL features of their databases. It positions itself as the Switzerland of data integration vendors.

There’s little argument that the ETL market is becoming ever more commoditized, as vendors such as Microsoft Corp. and Oracle Corp. continue to enhance the extraction and transformation capabilities of their relational database offerings, and BI vendors—such as Business Objects SA and SAS Institute Inc.—flesh out the ETL capabilities of their own offerings.

What’s surprising is that ETL stalwarts such as Informatica Corp. and Ascential Software Corp. don’t seem to be sweating the competition that much. We spoke last week with Sanjay Poonen, senior vice president of marketing for Informatica, who agreed that the ETL space is becoming more commoditized. In fact, Poonen says, Microsoft and Oracle will eventually control a sizeable chunk of the ETL tools market.

Poonen believes tools are only part of the issue: With so many different systems in the typical enterprise, he says, the data integration platform, and not the ETL tool itself, is the thing. In this regard, he says, there’s an even more important role for Informatica and others to play: That of the “Switzerland” of data integration offerings.

The ETL space is getting more crowded than ever. Microsoft, for example, is poised to release a substantially revamped Data Transformation Services (DTS) in Yukon [the next version of SQL Server]. Do you have any concerns about the moves of Microsoft and maybe others cutting in to your market share?

Well, it’s not just Microsoft—it’s sort of no surprise that every database vendor is strengthening the ways of being able to extract data out of their systems. IBM does it through partnerships, with us, with Ascential, and others, and Oracle has their own tool called OWB [Oracle Warehouse Builder]. OWB is positioned a little better in Gartner’s impression of their ETL capability than Microsoft is, by the way, so in response, Microsoft has had to do a lot to push their ETL solution--

But these tools are getting better with each release, aren’t they?

The tools market for ETL is where the database players are just going to get stronger and stronger and commoditize their tools, so you would expect over time that the capabilities that Oracle has with OWB and Microsoft has with DTS will get better. In fact it will commoditize ETL, so if people are extracting data out of SQL server, they’ll use DTS, and if they’re extracting out of oracle, OWB. In fact, we believe Microsoft will certainly give Oracle a run for their money here.

What about BI vendors, like Business Objects [which acquired ETL specialist Acta in 2002] who are getting into the mix, too?

Some of the other BI players that have front-end technology are getting involved, yes. Some of them acquired ETL tools that are very, very limited—for instance, [Business Objects and] Acta. We think those guys are very threatened by Microsoft and Oracle, because on the low end, all they’re doing is moving data from one source to one other source, so it’s mostly from one database to another. So with someone who is using Oracle, someone who is using SQL Server, they’re just going to say, “Why can’t I just use OWB or DTS to do that?”

You’re not worried about them saying the same thing if they’re existing Informatica customers?

Well, the key point here is that for both Oracle and Microsoft, these [ETL] tools are enhancing the furtherance of their platforms. But this is just the tools piece. It’s really the difference between [having] an integration platform, which is what we have, and [having] an ETL tool. There’s a sort of maturing process involved here, with the first generation being coding, the second generation maybe doing it a faster way than coding, the third being a platform where it’s used for all kinds of uses other than just data warehousing. The nature of a platform is different. It’s based on a very rich set of APIs so that people can build on top of that. We now have 800 APIs, we have 10,000 developers who build to these APIs. We have the most OEMs, we have 50 such OEMs.

A second criteria is the fact that the breath of connectivity to sources needs to go beyond the sources that come from one system. We joke about this internally that the diplomatic relationships between North and South Korea are warmer than between Oracle and Microsoft. It’s very hard for a company that focuses just on their data to be able to have a way where they can get data from outside their system. The challenge, I think, that DTS will have will be finding ways to get data effectively from not just IBM and Oracle, but also from the ERP systems, from other sources. Over time, that’s something that really has been our strength.

Also, it’s important to remember that typically the first generation of ETL tools—even the second generation—they’re typically used for data warehouses. Increasingly, we’re finding that a data warehouse platform has other uses, such as where the target the data is moving [to] is another operational system. This requires a much more robust integration platform.

In the low-end, where people just need ETL to extract data from one or just a few databases, where they don’t need an integration platform, per se, have you felt any pressure from SQL Server or Oracle?

At the low end of the ETL market where people are just loading data of one kind into SQL Server? I think there’s probably on the periphery, as any company grows, there’s probably a few customers that decide that they want to do it in a cheaper trend. They’re probably on the boundary cases. Our market focus is so much on the enterprise that we also don’t see a lot of these people. Microsoft’s focus is going to be the mid-size enterprise and the lower small- and medium-[enterprise]. We don’t focus very much on that market, and SQL Server itself hasn’t made a big dent into the enterprise.

About what percentage of your users is running SQL Server or Oracle?

We probably are about 20 to 25 percent of SQL Server customers, and about 40 percent of Oracle, and the rest are DB2. Most of our customers are using Oracle, but with both OWB and DTS, we’re finding that many people started out there when they had a very simple environment. It is our job to show them the complexity that they have. The research I think from analyst firms is that the average company has 10 or 12 systems, so as soon as you get the heterogeneity, the ETL tools [from Microsoft and Oracle] begin to lose their appeal.

Are you seeing cases where that’s happened? Where people get their feet wet using native ETL tools on a particular database platform, but then realize that they need something more?

Absolutely. There’s the halo effect of a large player that comes into the market and validates the market so to speak, and I think there’s probably some truth to that. I don’t have any numbers that I can give you, but I know that’s part of it.

Do you think there’s an opportunity to tap the capabilities of tools like DTS or OWB, so that maybe customers who start out using these tools but need something more, need to access these heterogeneous data sources, can continue to use these tools without disruption as they integrate more systems?

What we will do is this: SuperGlue [Informatica’s metadata repository] will extract metadata from those tools—that’s a natural case where in complex scenarios people are asking us to build metadata bridges. As you become a player, you start branching into other departments, and these other departments are using other ETL tools, and that’s really what our customers want, pulling the metadata out of all of these tools and getting a view of their lineage, so to speak.

So to wrap up, if you had to position Informatica’s integration offerings in relation to ETL tools from Microsoft, Oracle, and other database or BI players, how would you do it?

I would say that we’re the Switzerland of data integration [vendors]. We agree there’s a strong proposition that a lot of big players will make money in this, but I think there’s also a very strong proposition for Switzerland players like us, and it’s a very large market. It’s our focus as a company to be that trusted source of data to move data in a very heterogeneous environment. There will be people focused on the [departmental] silos [of data]—on SQL Server, on Oracle, on DB2—and there will also be those who are focused on being Switzerland players, like Informatica.

About the Author

Stephen Swoyer is a Nashville, TN-based freelance journalist who writes about technology.