Business intelligence (BI) has traditionally been used for supporting long-term strategic planning and short-term tactical tasks such as campaign management. Previous columns of mine have discussed how BI is now also being used to drive intraday business decisions and actions. This latter style of BI is sometimes called operational or real-time business intelligence.
Although the term real-time BI has strong marketing appeal, it has always been controversial from a technology perspective because the resource requirements of BI processing are a major stumbling block in a real-time environment. This is why I use the term operational right-time BI when discussing the use of BI in intraday business operations. Right-time represents the response time, or action time, required by businesses to react to a particular business situation. This action time may be a few seconds or a few hours, depending on business needs. Detecting fraud, for example, requires a faster action time than determining that a store has run out of a particular product.
Action time can be broken down into three components: the time required to capture business transaction data for analysis, the time required to analyze this data and deliver the results to the user, and the time for the user to make a decision and take action.1
BI vendors have improved action times by optimizing the tasks involved in producing BI for decision making and action taking. Data capture times have been reduced by replacing batch extract, transform and load (ETL) jobs with data capture processes that stream data changes and events into an operational data store (ODS) or data warehouse. To help reduce action time even further, some of the new operational BI performance management products appearing on the market can analyze stream and event data in-flight, without the need to first load it into a physical data store such as an ODS. The BI analysis component of action time has been improved by database management system (DBMS) capabilities such as parallel processing, advanced analytical processing engines and materialized views. Business user reaction times have been reduced through the use of alerts, guided analysis and automated recommendations.
All of the BI processing discussed thus far occurs asynchronously in the background while business transaction applications continue to run the business. IT organizations have always been reluctant to allow BI processes to interfere with the strict service level requirements of business transactions. This position is changing, however, as companies have begun to realize that fast action times for solving business problems and meeting customer needs can offer significant competitive advantage. This change in attitude by IT has, in turn, led not only BI vendors to focus on the use and integration of BI in daily business operations, but also the suppliers of business transaction (BTx) processing solutions.
To date, most of the focus on analytical processing by the BTx software vendor community has been in the area of business process management. This software not only analyzes and automates business processes, but also monitors them. Products such as IBM WebSphere Business Integration Monitor and TIBCO BusinessFactor provide business analytics about the performance of business processes and underlying business transactions. These products can also retrieve historical data from a data warehouse to put the performance analytics into a business context. As with BI products, this analysis is done asynchronously from the BTx processing.
We are likely to see increasingly more support from the BTx software vendors for BI processing. IBM's recently acquired Alphablox product, for example, will be used to enhance the analytical processing capabilities of the WebSphere Business Integration Monitor. TIBCO has announced OpsFactor, which analyzes events in business process flows created using its BusinessWorks product.
The most important new facility coming from BTx software vendors is most likely the ability to embed BI processing in the BTx process flow itself. This can be thought of as embedded or in-line BI processing. This type of BI processing is particularly useful in a service-oriented architecture employing technology such as Web services. The required BI functionality can be defined as a Web service and then called from the business process flow as required. Examples of BI Web services here include customer segmentation and fraud detection. Of course, BI performance is still an issue here, and in most cases, the BI Web service would employ cached analytics, stored data mining models and so forth.
In addition to invoking BI Web services, in-line BI processes can also issue SQL or XQuery statements to retrieve data via enterprise information integration (EII) middleware. One impressive demonstration of this that I recently saw involved the use of BEA's Liquid Data EII software to connect to a Hyperion BI environment. The BEA application in the demonstration used Liquid Data not only to create a planning dashboard of strategic and tactical BI data, but also to assess the impact of operational events on the business. If a significant procurement order is delayed, for example, information about the financial impact of the delayed order is sent in an alert to an operational dashboard. The operational dashboard, in turn, advises the business user about the types of BI reports that can be run to perform more detailed analyses of the issue. This is a clear example of how in-line BI can be used to improve the capabilities of the business process.
We can see then that BI can be used in many different ways in an operational environment. In some cases this involves the use of operational BI and performance management products to provide right-time analytics, and in other cases this may entail the use of in-line BI to support close to real-time processing. It is also clear that the trend toward the use of operational BI will lead to BTx software vendors competing with BI vendors for this slice of the market.
1. Hackathorn, Richard. "Reducing Action Distance." The Data Administration Newsletter, February 2004.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access