When IBM said in March 2005 that it would be acquiring data integration software vendor Ascential for $1.1 billion, the question market watchers were asking was not so much ‘why has IBM bought Ascential?' as ‘why did it wait so long?'
As Ian Wesley, an analyst at technology market watcher Ovum, observed: "This acquisition was long overdue. It makes so much sense for both parties that one can only assume that Ascential had been holding out on price."
Certainly IBM could have made its move into Ascential's area of data quality, migration and integration a lot earlier. In fact this is the second time IBM has sat down with CEO Peter Gyenes and the Ascential board.
Back in April 2001, IBM bought the Informix database business, but left standing Informix's data integration arm that had originally been part of Advent Software. Flush with the $1 billion cash it took from IBM, Gyenes renamed the remaining business Ascential.
But IBM did not go away. Instead, it forged a close partnership with Ascential, reselling its tools for extracting and cleaning data from multiple sources.
In the four years since that deal, IBM seems to have reassessed the importance of having its own data integration technology – and for multiple reasons. First, the data integration market has finally taken off. After years in which extraction, transformation and loading (ETL) software companies struggled, strong demand is now being fuelled by issues such as regulatory compliance and the pace of acquisitions. As a result, companies such as Ascential, Informatica, Sunopsis and others have delivered a stream of impressive performances.
And that looks sustainable. Analysts at market watcher IDC predict that worldwide data integration spending will rise from $9.3 billion in 2003 to $13.6 billion in 2008.
IBM's EMEA marketing manager for information management, Angus Faulconer, puts that growth down to the shifting priorities of IT managers. "Flexibility [of information management] is the biggest concern among CIOs. Information management is now absorbing 40% of IT budgets," he says.
The second reason for IBM's move centres on the way in which applications can now communicate with each other using web services standards. That has brought application integration and data integration much closer.
Application integration has been a key part of IBM's WebSphere middleware portfolio since it acquired Crossworlds in 2002. It has seen a growing requirement by customers involved in large application integration projects for data integration tools as well. Reacting to that, data integration companies have been moving in the opposite direction – though not always successfully.
"Application and data integration are part of the same problem," says Aiaz Kazi, general manager of enterprise backbone and business integration at Tibco.
With that in mind, Ascential bought applications integration specialist Mercator for $106 million in July 2003 – but the technology has since lost momentum in the market against other integration vendors such as WebMethods, Tibco and BEA. Nonetheless, getting at that Mercator legacy base is another reason IBM was keen to open its wallet.
The Ascential acquisition puts pressure on IBM's competitors to match its play. Tibco recently expanded its offerings by including data integration technology licensed from Embarcadero in its BusinessWorks integration suite.
Kazi thinks that by pricing the data integration as part of a combined product Tibco will be able to establish itself as a complete integration software provider and fuel growth of its applications integration and business process management software.
"We will sell data integration along with application integration, but at a lower price point," says Kazi. "The traditional ETL market would sell you a toolset for $400,000 – we'll sell you the same or better at $50,000."
It is fighting talk, of course, but Tibco's Kazi is already drafting the obituary for some vendors. "The standalone ETL market," he says, "is dead."