Has IBM Bitten Off More Than It Can Integrate?

By Lisa Vaas  |  Posted 2005-05-04

Has IBM Bitten Off More Than It Can Integrate?

WESTBORO, Mass.—Like a cherry on top of a sundae, IBMs Information Management division on Wednesday put its newly acquired Ascential data integration technology on top of a stack of acquisitions that, it says, is deeper and broader than any existing content management technology.

"For the last 40 years, the IT industry and database companies have been fixated down here, at the data level, [concentrating on] storing data efficiently and cost-effectively, backing it up, archiving it, making sure theres integrity with transactional data," said Janet Perna, IBMs general manager of Information Management.

Perna indicated a slide that displayed layers of storage, data repositories such as databases, applications, analytics and other layers of the cake that is enterprise infrastructure.

"But pretty much it had no value once created," Perna said. "I call it passive data, or you can think of it as dead data. Most companies are not getting tremendous value from that data. Value comes from taking the data and integrating it with other pieces of information that will enhance it and enhance the business value."

Perna was speaking during a press event to roll out the first fruits of IBMs Ascential acquisition. The event featured a panel of customers and data integration experts at Ascentials headquarters here.

Also in attendance were Pete Fiore, former Ascential president and now vice president of Information Integration Solutions for IBM, and IBM Distinguished Engineer and Vice President of Strategy for Information Integration Nelson Mattos.

Creating value out of Pernas so-called dead data has been the motive behind a steady stream of IBM acquisitions:

Trigo, for the integration of product information; Aptrix, Tarian and Green Pasture, all for content control; Alphablox and SRD for real-time, embedded analytics; CrossAccess and Venetica for federation of non-IBM database data and federation of content, respectively; and IBMs most recent acquisition, Ascential, for high-speed ETL (extraction, transformation and loading), data profiling, data cleansing and metadata management.

Why so many pieces to this data integration puzzle? Because the reality of knitting together systems in order to get to IBMs On-Demand nirvana is extremely difficult to do, Perna said.

Read more here about IBMs on-demand strategy.

"Information is all over the place: Its in silos of applications, its in different formats, different databases, different content repositories," she said. "Eighty-five percent of business information is … semistructured, in imaging systems, on the Web, in file systems, in e-mail. Its all over the place. A lot of time is spent, and money is spent, in companies, just managing this information. Forty percent of peoples time is spent integrating information."

The grand challenge, Perna said, is helping companies to simplify this environment, so companies can use this data, put it into context, analyze it and use it to provide business insight.

"That will require an end-to-end information infrastructure that enables companies to deal with all this information," she said. "That starts with infrastructure down at the physical layer, the physical data assets, the storage area networks that allow us to virtualize data.

"Then there are higher-level services that help deal with How do I find it, how do I analyze it? That requires information integration, search and access technologies, analytical technologies, as well as content control to do archiving, electronic records management, etc.

"Were even seeing extensions like master data management to be able to provide a unified view for customers, citizens, suppliers or what have you," Perna said. "The investments IBM is making in information management are addressing this spectrum: from physical storage, from Tivoli storage and Tivoli storage management, to DB2 and Cloudscape, up to technologies around master data management.

"The acquisitions weve been making are not helter-skelter here. Were filling in capabilities through internal investment or through acquisitions were making as we fill this out."

Click here to read about IBMs new wireless education services.

Still, questions remain. As IBM reaches to give customers integrated information, with Google-like enterprise search, the ability to create relationships between scanned images (unstructured data) and customer records (structured data), and all else that the term promises, is it going to make enterprise lives simpler or more complex?

Tony Baer, an analyst for onStrategies, said that, for end users, all the acquisitions shouldnt matter. "All they care about is I want my data now," he said.

Next Page: It matters what lies beneath.

What Lies Beneath

But for IT managers, whats beneath the pretty search engine matters—a lot. "For IT folks, it means, as with any product, you have to get to know the underlying architecture of the product youre getting and what level of integration" it will have, Baer said. "Will the integration be at the metadata level? If so, how detailed is the metadata? Does it come under a common shell, or do you have to go out to different shells under different programs?"

This is important. The more things that are common, the fewer things you have to learn. If all products acquired by IBM come in under a common shell, its a lot easier to learn than different programs, Baer said. If its a federation of programs, sharing a common interface but running different engines under each product, life is hard.

Mattos addressed this point during a question and answer session following Pernas presentation and a panel presentation of customers and data integration experts.

"[The portfolio] is integrated in many different places," he said. "At the metadata level, at run-time with autonomic capabilities … Its integrated in the end user point of view: one for the business analyst and one for the application developer, etc. However, our vision is not to have full integration at the run-time, because that would require the customer to rip and replace the run-time that they have."

In other words, Mattos said, IBMs solution trumps that of Oracle Corp. in that IBM lets data reside wherever it wants to. As it is, the next version of WebSphere Information Integrator, code-named Hawk, will deliver improved integration. Of particular note is the ability to share metadata between the suites tools, thus eliminating redundant metadata discovery.

Still, Mattos said, the Information Management division is working toward common capabilities across its portfolio. "One of the things weve heard today is about increased investment in research and development," Mattos said. "As a result, not only are we delivering the Hawk release, we also added capabilities in both releases to improve integration of both products.

"The initiative around Autonomic Computing started with DB2," Mattos said. "When I think about [Pernas] vision… the vision we have, the information infrastructure becomes one thing. The picture I showed of information infrastructure is one layer that simplifies all of the information assets companies deal with. What I would like, at some time, is the ability to administer this environment with a common set of capabilities. Today if you look at Content Manager … its a little different from DB2 capabilities, which are a little different from Information Integrator capabilities. What Id like to see is bringing all the pieces I have together."

Speaking of bringing pieces together, who exactly owns the new products that IBM keeps acquiring? Is the Ascential technology under the DB2 brand or the WebSphere brand? IBM has labeled it WebSphere, but Pernas heading-up of the effort shouts "DB2."

Perna shook her head and laughed as she tackled that question. Why? Its a question that IBM gets often.

Thats because, up until recently, IBM as a company was as siloed as the disparate content repositories its attempting to integrate with its current crop of tools. "In the old days, each area of IBM would be a separate universe," said Judith Hurwitz, an analyst with Hurwitz Associates and the moderator of IBMs customer and expert panel. "Youd see one group competing with another."

Luckily for customers, thats changed. "Whats changed over the last three years is each unit provides services to each other," Hurwitz said. "Youre seeing more transfer of knowledge and capability and product between groups. Theres a lot more sharing across business units."

Thats important for customers who need to know which throat to choke. Nowadays, Perna said, brands and executive capabilities are blurring together. Wednesdays product news is WebSphere, its business integration, its all things information, and if its information, it resides with Perna.

Lisa Vaas is Ziff Davis Internets news editor in charge of operations, aka Conference Goddess. She is also the editor of eWEEK.coms Database and Business Intelligence topic center. She has been with eWEEK and eWEEK.com since 1995, most recently covering enterprise applications and database technology.

Check out eWEEK.coms for the latest database news, reviews and analysis.

Rocket Fuel