A Data Mart for the Internet - Company Operations

QueryObject Systems CEO says business intelligence needs to focus on the unknowns.

Q: Can you describe the evolution of business intelligence tools?

A: The Executive Information System in the late '80s found ways to distill information about business hidden in data byproducts and deliver it to a few people at the top of a hierarchical business organization. In the early '90s, along with downsizing and the flattening of the hierarchy, client/server technology arrived. The ability to take this concept of analyzing data off the server and provide it at departmental-level views, courtesy of client/server, is what drove business intelligence. Today, people are gearing up their BI systems to take advantage of the Internet, and the continued pushing of decision-making out to the expanding boundaries of the organization.

Q: How has the BI user changed?

A: The premise in BI was we'd deliver a specific set of data to a predefined set of users who use tools that we know about in advance, to perform a kind of analysis that they had defined in advance. In the new business environment, it's not going to be possible to determine in advance who the users are going to be, because organizations are going to expand and collapse based on the tasks at hand. We're not going to be able to determine in advance what they're going to do with the information, because we can't anticipate all of the tasks. And we won't be able to predetermine what kind of tools they will want to use.

So the new class of application is going to demand that there is a very large and heterogeneous set of data available on an ad hoc basis, to a wide number of potential users, who will be using it to do a variety of types of analysis using a variety of tools. While BI is focused on the knowns, this New World is going to have to deal with a number of unknowns, and that puts a different set of constraints on the tools.

Q: Can you give me an example?

A: Number one, you've got to have a lot of data because you can't predetermine what you're going to give people; two, you've got to make it available in a polymorphic format because you don't know what they're going to do with it; three, it has to be in an open system because you can't anticipate whether they're using Brio Business Objects, Excel spreadsheets, or something we don't know about yet; and four, the sheer potential number of users means we probably have to go to a distributed multitier architecture rather than today's server-centric architecture.

Q: What are other BI megatrends?

A: Rather than looking at data output from a simple system and just managing that area, a knowledge worker may be responsible for a much broader view of the business. So certainly data volumes are an issue. Combine that with the astronomical growth in data itself as we move business over to the Internet. Data volume is the Achilles' heel of business intelligence, and is one of the must-solve issues to move forward into the infrastructure of the information economy.

Q: What is the QueryObject System?

A: The QOS is a data mart designed specifically for the Internet economy; it can provide both a relational and a multidimensional view of data, and it can manage extremely large volumes of source data. It does not come with its own analysis component; it speaks through ODBC, JDBC and OLE/dB for OLAP to the plethora of tools out there. Most significantly, it not only supports today's server-centric BI Internet architectures, it also has a new feature, Distributed Replication Server, which streams the analytical data across the Internet to support multitier applications. Traditional BI tools view the Internet as a communications vehicle for distributing reports. The QueryObject view is that the Internet is a large network computer, and we should be providing a networked data model that allows analysis to happen at a multiple number of nodes across the network.

Q: What is the advantage of the Distributed Replication Server?

A: The first advantage deals with the scalability of user counts. Rather than having a number of users banging away at the same server, people have taken the data off and are using it on their local analytical node. So potentially you can scale the system to almost any length because you no longer have to worry about managing concurrency. As a byproduct of that ability to data stream, one of the things we've been able to do is Web enable the BI power user desktop. The choice you have to make with BI tools is, do I take the high-performance desktop version or do I use the limited-functionality Web version? We can stream data from the host QueryObject System into one of the power user desktops--the tool functions exactly the same and believes it's working with local data, but it's extracting data on demand over the Internet.

Q: Will the BI market go away?

A: I'm predicting that BI needs to recognize that the problems it's solving today are not the problems it has to solve tomorrow.

Robert Thompson is chairman and CEO of QueryObject Systems Corp., developer of scalable Internet analytical technologies. He previously spent eight years in corporate and marketing positions at Cognos Corp., a longtime player in the business intelligence (BI) market. Thompson spoke with Managing Editor Colleen Frye about current trends in the BI market and the Internet economy.

Quick Profile: OueryObject Systems Corp.

Founded: March 1998, reincorporated from the consulting firm CrossZ Software

Headquarters: Roslyn, N.Y.

Publicly Traded: American Stock Exchange and Boston Stock Exchange

Revenues: $1.022m (6 months ended 6/30/00)

Growth From Previous Year: 69%

Employees: 60 Customers: 20+

COPYRIGHT 2000 Wiesner Publications, Inc.
COPYRIGHT 2001 Gale Group