In-Depth

Steady Is the Course for ETL Pure Plays

DataMirror and SunOpsis continue to find their own way—a third way, so to speak—in a hotly contested ETL marketscape

• By Stephen Swoyer
• 12/07/2005

ETL doesn’t begin and end with Informatica Corp. and IBM Corp. There are still a host of ETL pure plays, after all, and—by the looks of it, anyway—most of them are doing just fine, thank you. Two textbook examples are DataMirror Corp. and SunOpsis Inc., which continue to find their own way—a third way, so to speak—in a hotly contested ETL marketscape.

Last month, SunOpsis proved that there’s still plenty of demand for third-party ETL solutions, notching an almost 70 percent year-over-year increase in revenues. In a statement released last month, the company attributed this performance to the expansion of its operations in the EU, along with the acquisition of several large customers—including Overstock.com, telecom giant Caudwell Communications, and global shipping specialist ANL Container Line.

DataMirror, for its part, had an eventful November. Just before Thanksgiving, for example, the Toronto-based company reported its Q3 earnings (for Fiscal Year 2006). Then, last week, DataMirror announced version 5.3 of its flagship data integration suite, Transformation Server.

For most of its history, DataMirror’s fortunes have been yoked to those of IBM’s could’ve-should’ve-but-for-some-reason-never-was a juggernaut iSeries (nee AS/400) midrange platform. As iSeries marketshare declined, so, too, did DataMirror’s growth. Over the first few quarters of 2005, for example, DataMirror barely grew its revenues in year-over-year comparisons. And DataMirror’s earnings for the first three quarters of its 2006 fiscal year actually trail last year’s performance: Year-to-date net income ($2.2 million) is about forty percent down from last year’s tally ($3.5 million).

That comparison is deceptive, however. Fact is, DataMirror’s FY 2005 earnings were bolstered by the sale of a $4.1 million investment that helped to generate $3.9 million in income in the second quarter of FY 2005. If anything, with three consecutive quarters of positive revenue and income growth behind it, the DataMirror of 2006 is outperforming the DataMirror of 2005.

To some extent, this rebound mirrors the iSeries’ own (even more astonishing) rebirth: In October, for example, Big Blue reported that iSeries revenues had increased by 25 percent from the same period last year. And through the first few quarters of 2005, IBM’s iSeries platform also managed positive growth – 1 percent in Q1 and 10 percent in Q2. It’s encouraging news, but at the same time, DataMirror’s prospects are likely to remain limited—at least to the extent that it’s still perceived as a mostly-iSeries play. Indeed, Gartner Inc.’s latest market research data suggests that DataMirror’s ETL marketshare contracted again last year, slipping by 3.7 percent.

That brings us to Transformation Server 5.3, which is quite possibly DataMirror’s most mainstream-oriented data integration offering to date. In addition to DataMirror’s bread-and-butter ETL and data integration capabilities, the revamped 5.3 release includes Master Data Management (MDM) capabilities and Web Services support, and also ships with DataMirror’s Change Data Capture (CDC) technology, which officials position as a superior alternative to trigger-based or table scan-based approaches. Transformation Server 5.3 is also a cross-platform play, with support for z/OS, AIX, Solaris, HP UNIX, Linux, and Windows—in addition to OS/400, of course.

The new release ships with a number of new goodies, including conflict resolution capabilities that promise to reduce or eliminate conflicts that occur during bi-directional replication; complex transformation capability, which enables real-time data consolidation for operational reporting and dashboarding activities. Transformation Server 5.3 includes several usability enhancements, including a wizard-based interface that allows users to configure audit processes and create audit trails; new start/stop functionality that lets administrators initiate or kill data replication projects directly from within the monitor or from the Enterprise Administrator GUI; one-click project management; latency monitoring and alerting, which provides a more accurate picture of the health of data integration activities. This last feature can also alert administrators when processes operate above pre-defined latency thresholds.

For a long time now, DataMirror’s iSeries success has been a curse as well as a blessing. (To date, DataMirror and SPSS’ Inc. Showcase technologies are the most widely used data integration solutions for iSeries.) While the iSeries is an exceptional platform in its own right—and while iSeries license renewals comprise a healthy chunk of DataMirror’s maintenance revenue stream—DataMirror’s association with that platform often functioned as a kind of impediment to breakthrough (or, at least, mainstream) success.

DataMirror has long struggled to position itself as an other-than-iSeries play, however, perhaps most notably with the Operational Data Store (ODS) offering it announced last year. (http://www.tdwi.org/Research/display.aspx?ID=7572) ODSes were hot then and are hotter now, and there’s a chance DataMirror’s cross-platform-friendly Transformation Server 5.3—which ships with core ODS capabilities (e.g., ETL and CDC), along with CDC-enabled Master Data Management features—could gain some ground.