ETML Tools Vendors

In a prior column ("Requirements for ETML Tools," July 15), I discussed some of the key issues that you as a developer need to consider if you're going to select a tool to manage the extract, transform, move, load (ETML) process to populate a decision-support database.

The biggest competitor of the ETML tools is the in-house programmer. Many shops think they are saving money and getting better value if they write their own routines to perform the functions required to prepare data for the decision-support database. While this may be true for a simple set of uncomplicated extractions that will feed a single data warehouse, it is less true in an environment with multiple source databases, complicated transformation requirements or multiple target decision- support databases.

When you're evaluating a tool, be sure to consider the life-cycle cost of the tool, not just the price. A tool with a relatively high list price may actually be more cost-effective than a cheaper alternative. Now, let's talk about the various vendors.

Ardent Software Inc. (Westboro, Mass., -- DataStage from Ardent Software runs on Windows NT, and uses EDA/SQL from Information Builders Inc. (New York) to extract data from a wide variety of legacy systems. Developers can extend a very complete set of transformation functions by using Visual Basic. Ardent recently signed partnership deals with Sybase Inc. and Informix Software Inc.

D2K (San Jose, Calif., -- D2K’s Tapestry has strong integration with multiple data sources, including SAP and PeopleSoft. It includes Tuxedo, a transaction monitor from BEA, to ensure that the ETML process can be restarted at the most recent checkpoint in the event of a failure.

Informatica Corp. (Menlo Park, Calif., -- Over the past year Informatica has released a new architecture to support enterprisewide data mart deployment. Consequently, Informatica has strong metadata repository technology, along with a rich set of transformation functions. The ability to extract data from arcane or mainframe data sources is limited, requiring either an ODBC interface or flat file to get at the data. The scripting language is a proprietary variant of SQL.

Platinum Technology (Oakbrook Terrace, Ill., -- Platinum Technology has been in this business for a long time, and has excellent mainframe connectivity, a complete set of transformation functions, and high-speed data movement capabilities. Recently, Platinum Technology integrated these tools into DecisionBase. The scripting language, unfortunately, is proprietary. However, this is offset by Platinum's repository partnership with Microsoft Corp.

Prism Solutions (Sunnyvale, Calif., -- Prism has also been doing ETML for a long time. This year the company re-engineered its Executive Warehouse to make it more flexible, and to enable it to generate extraction and transformation code in several languages, including COBOL, C/C++, Java and SAP’s ABAP. The product is strongly oriented toward the mainframe environment, and used mostly in high-end environments.

Sagent Technology (Palo Alto, Calif., -- Sagent provides a complete NT data mart solution. The developer user interface is nicely done, and the software has a full range of transformation functions. However, if you elect to use Sagent to build the data warehouse, you must use Sagent on the client side, as well.

Other vendors that you might consider, but that we don't have space to review here, include Carleton Corp. (Eden Prairie, Minn.,, Constellar Corp. (Redwood Shores, Calif.,, DataMirror Corp. (Markham, Ontario,, Evoke Software (formerly Dbstar Inc., San Francisco,, Group1 Software Inc. (Lanham, Md.,, Praxis Int'l Inc. (Waltham, Mass., and Red Brick Systems Inc. (Los Gatos, Calif.,

A note about decision-support database updates. These vendors, with the exception of Platinum and Sagent, support one form or another of change data capture. This is an important consideration if you want to be able to add only new or updated data to the data warehouse. Otherwise you'll be forced to completely repopulate the decision-support database every time you refresh it.

All of these vendors provided customer references to me, and the references were uniformly positive. This reflects two things: Vendors usually make an effort to find customers who will act as good references, and people who are willing to discuss their projects with analysts or the press generally have had a positive experience, while people who failed are not usually willing to talk to outsiders. An excellent source for relatively unbiased references that aren't affected by either of these influences is the comp.databases.olap newsgroup, which is the closest thing to a decision-support newsgroup on the Web. -- Robert Craig is director, Data Warehousing and Business Intelligence Division, at Hurwitz Group Inc. (Framingham, Mass.). Contact him at or via the Web at