Breaking Out of the WarehouseTo many, BI is equivalent to data warehousing. Although data warehouses are important, they're not the total solution. To deliver all the BI capabilities business users want, we need to create a BI gridWhy do executives still lack consistent, reliable access to information, even as the technology for BI applications advances and the techniques mature? It seems that our vision of BI continues to outdistance our ability to make it materialize. You would think that as our technology improves, so should our ability to gather, synthesize, internalize, and exploit the data streaming into our organizations. But rarely does this assumption play out in reality. BI projects throw many challenges our way, including the volume of incoming data, the complexity of data integration required in today's global enterprise, and the users' insatiable appetite for information. While these challenges continue to hamper some BI efforts, they still don't account for the entire gap between the available tools and techniques and the analytic solutions being delivered. Furthermore, it simply is not acceptable to say that BI is unachievable because there's too much data, it requires complex integration, or users are never satisfied. The purpose of BI initiatives is to take that unwieldy, dynamic enterprisewide data and turn it into a predictable stream of decision-support content that addresses the requirements of our user communities. But, our ability to exploit the massive amounts of data our organizations generate isn't resolved by the traditional technology we introduce or the common techniques we attempt to implement. Worse still, the more technology we add, the more fragmented our information becomes — with disparate databases and intricate interdependencies. Warehouse BiasOne major reason the unbridged gap exists between the promise and reality of BI is the way in which BI initiatives are planned and conducted. Contemporary thought squarely focuses BI on a foundation of data warehouse technology and techniques. Consequently, many BI initiatives offer little more than warehouse-centric solutions. Don't misunderstand: Data warehousing is vital. It ensures data quality, historical integrity, and completeness of record — all of which are critical for strategic analysis. The data warehousing process includes periodically gathering disparate data, cleansing and integrating that data, and then loading it into static, persistent data structures. A data flow is established from operational sources to a staging area, then to the atomic-level warehouse, and finally to data marts. When implemented correctly, this process provides a single version of the truth across the enterprise as well as the basis for strategic analysis. For that reason, the warehouse architecture serves a critical role in BI environments. But the value of warehouse-centric implementations is also their limitation. The data flow from one physical structure to the next requires a batch window. Data warehousing also dictates the type of technology necessary to implement the process. The data warehouse architecture constrains the types of BI applications possible and, consequently, a majority of business requirements for BI go unsolved or underserved. The right approach to filling the BI gap is to complement the warehouse toolset with agent-based technology. This software is specifically designed to reach across the layers of the BI environment, essentially covering the enterprise in a grid or fabric of agents. Each agent is capable of monitoring, collecting, and analyzing data and events from virtually any object in the environment. Moreover, a composite or correlated view can be created from a predefined set of data or events, greatly increasing the decision-guiding power of BI.
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| |||||||||||||||||||||||||||||||
