IBM and Search
What's up with IBM and Enterprise Search?
by Stephen Arnold
03-Nov-2005
[Editor's Note: in the wake of IBM's acquisition of niche search vendor iPhrase as well as the joint announcement that Google will be providing desktop search for the IBM's OmniFind enterprise search system, we asked CMS Watch Search Analyst Steve Arnold to put together a quick FAQ on the bigger picture.]
How does iPhrase "fit" with DB2 search, OmniFind, WebFountain, and other search "systems" available from IBM?
iPhrase, as we understand it, will become part of the WebSphere offering. iPhrase converts unstructured inputs into structured XML instances. When a document is retrieved from the iPhrase repository, the document is reconstructed. There can be visual differences from the original and the iPhrase reconstruction, but they are for the most part minor. iPhrase, therefore, adds zip to the current WebSphere line up as well as creating a demand for more storage, DB2 licenses, reseller customization, and other "value adds" from today's IBM.
How does iPhrase fit with the Google desktop search?
There is no exact fit. Google can, in theory, see any content properly exposed to it. iPhrase is a repository-based search system with powerful metatag features. The likely mix will be to allow Google to see all or part of the iPhrase data via an adaptor created by IBM for its resellers to further customize.
What about OmniFind?
OmniFind searches unstructured information in a DB2 environment. Some of OmniFind's features are unique to DB2. If iPhrase content is in a DB2 structure, then iPhrase can make use of OmniFind functions. Customization--which is the lifeblood of IBM-certified resellers--will be needed.
What about WebFountain?
WebFountain is a toolkit with a crawler. An IBM customer can avail itself of the WebFountain content acquisition, parsing, and report functions. Customization is required along with integration, storage, etc.
What's the bottom-line on IBM's strategy in search?
Sell licenses to the bits and pieces that must be glued together to build the system that an IBM "true blue" customer requires. IBM does not sell applications. So, IBM's approach to search is to make certain that whatever its customers require, IBM has a box of parts from which an F1-type search and retrieval system can be assembled.
Is the tie up between IBM and Google an anti-Microsoft play?
As Groucho Marx said, "Your bet your life." IBM has not forgotten the 20 years from 1984 to 2004 when Messrs. Gates and Allen (then Mr. Ballmer) forced IBM to become a computer company with an unbreakable addiction to consulting revenue. IBM and Google, with Sun Microsystems, represent a triumvirate of capability that will try to make Windows "Live" into Windows "Dead."
How many search systems does IBM really need?
As many as it takes to remain one of the leading hardware, software, and services firms in the world. Remember STAIRS III? Well, IBM has more than a dozen search "systems" in its toolkit, not including third parties who are authorized vendors. IBM has been serious about search since the mid-1960s, and it will remain serious for the lifetime of the company. Remember Clever? That's an IBM widget that influenced Google (Messrs. Brin and Page acknowledge the technology in their writings about search). IBM may be shine with fewer lumens than some companies, but the folks in White Plains, Armonk, and Almaden have search on their minds, patents in their pockets, customers in their ledgers, and Microsoft in their gun sites.