Search Tools Not a Substitute for Data Organization

Ann All

Much recent ink has spilled about the expected dust-up between Microsoft and Google over control of the online advertising market, thanks to Microsoft's attempt to block Google's purchase of DoubleClick and Microsoft's own efforts to buy Yahoo.

 

And the continuing speculation that Google Apps will some day challenge Microsoft's Office (or at least Office Live) makes for an endlessly entertaining story.

 

But the biggest battle between the two may be brewing in the search market. I know what you're thinking, Google has that all locked up. That may be true in the consumer space, as it's tough to imagine how any company could challenge Google there now.

 

No, I am referring to enterprise search, an entirely different animal from the consumer version and a market Google has yet to convincingly crack. In fact, industry observers like CMS Watch have found Google's enterprise search offerings, notably its Search Appliance, do not offer as much functionality as similar products from other vendors.

 

With its purchase earlier this year of Fast Search and Transfer, Microsoft appears to now have all of the enterprise search bases covered: free or low-cost entry-level product; a mid-range offering, that will likely be used by most of its customers; and an ultra-sophisticated and ultra-expensive option.


 

ZDNet's George Ou recently wrote that Microsoft's low-end offering, Search Server 2008 Express, seems like a "must download for any Windows Server shop." Even without the latest versions of Windows Server, Ou says it offers "a great way to build a very cheap enterprise search engine appliance with a minimal Windows Server 2003 or above license and a simple 1U server for less than $2,000" -- which he points out is far less than what Google charges for similar functionality.

 

Interestingly, the comment string following Ou's blog post ventured off into what seems like a somewhat misguided debate over data organization, with some readers apparently seeing Ou's endorsement of the product as a tacit admission that they shouldn't bother trying to organize their company's information. Wrote one:

Possibly nice if ... 1) You're a Microsoft only shop. 2) You don't insist (or at least try) to keep your company's documents organized. Search (in the Google or any other sense) is the refuge of a disorganized mind, or a information problem that's gotten out of hand.

Ou responds that no matter how well organized your data, an indexed search product will make it easier to find. "Humans simply aren't good at finding information manually," he writes.

 

Another reader chimes in that no search tool will be effective without well-organized data. Relaxing data organization requirements (as readers seem to think will happen with Search Server 2008 Express or similar products) creates "a bigger nightmare for your sysops" because they need to keep tabs on what information can be found where. There is a fair amount of impassioned back-and-forth on this issue, with Ou repeatedly insisting that he didn't mean that search is a substitute for well-organized data.

 

The fact is, organizations are increasingly seeking the ability to locate both structured and unstructured data, especially in relation to their business intelligence efforts. In a January IT Business Edge interview, Ventana Research's David Stodder credits (you guessed it) Google with changing expectations about enterprise search. Stodder said:

There's definitely a demand for better search and the ability to locate information that isn't as structured as a relational database. ... The kinds of tools we have for BI really are not appropriate for looking for that kind of information, at least not directly. That's one issue. The other issue is that people are interested in getting information quickly. The approach that a BI tool takes is you write a query and send it to the system. It takes a while to get the answers. So with people sort of spoiled with Google, you can type anything into its search engine and you're going to get an answer back. It might not be the answer you're looking for, but you're going to get some kind of answer back in seconds. It's a matter of figuring out how search could get people information much more quickly.

Another issue, as I see it, is that search tools are going to become increasingly essential as the amount of information gathered and stored by organizations grows. (And grows -- check out these alarming statistics on data growth from a recent IDC report.)



Add Comment      Leave a comment on this blog post
Mar 13, 2008 1:48 AM George Ou George Ou  says:
Thanks for linking to me. I would like to point out that I said that neither is a substitute for the other, you need both organization and search on the desktop and in the enterprise.I had been a doubter of indexed search until I started using it on my desktop in Vista. I realized that my usage of search was sparse before I had indexed search because my mind had already factored in the painful time it takes to find something. After I had indexed search, my usage of search went up 100 fold. Reply
Mar 14, 2008 2:38 AM Kingsley Idehen Kingsley Idehen  says:
Enterprise search continues to be mangled with Enterprise Query and/or Enterprise Data Access.An organization wants to be able to explore relationships across clearly defined entities across a broad range of domains. They want to be able to construct agless entities such as: Customers, Orders, Suppliers, Invoices, Employees etc., across diverse data sources (enterprise databases, xml databases, web services etc). The Linked Data enclave, within the Semantic Web realm , is where the real answers lie to what is inaccurately described as "Enterprise Search".Here is a live example of the popular Northwind Database (used with ACCESS, SQL Server, and the recen Astoria Data Services from Microsoft), that uses the Web's HTTP protocol to fashion a Distributed Data Object system (via ID/URIs) that makes relationship discovery and traversal a matter of mouse clicking i.e. real information at you fingertips, the kind envisioned long befor the Full Text Search distraction that is "Enterprise Full Text Search".Here is a live example of the Data Graph reveal by simply de-referencing (via HTTP) the Data Object identified (via a URI) as :http://dataviewer.zitgist.com/?uri=http%3A//demo.openlinksw.com/Northwind/Customer/ALFKINote: You can explore the data graph by clicking on the Data Links which expose Web Pages that reveal descriptive Attributes and Relationships across the entities in the hypothetical Northwind Enterprise.Final note, this is all achieved via OpenLink Virtuoso (http://virtuoso.openlinksw.com). Reply

Post a comment

 

 

 

 


(Maximum characters: 1200). You have 1200 characters left.

 

null
null

 

Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.