Perhaps the biggest advance I have seen over the past year is that companies are beginning to understand that there are many different types of real-time processing, each with its own uses, benefits and issues. IT staff are also realizing that right time is a better term to use than real time. Right time implies that different business situations and events require different response or action times. When planning a right-time processing environment, it is important to match technology requirements to the actual action times required by the business - some situations require close to a real-time action, whereas with others, a delay of a few minutes or hours is acceptable.
There is a wide variety of definitions for the different types of right-time BI processing that exist, but most real-time BI requirements can be placed into one of four main categories: right-time data integration, right-time data reporting, right-time performance management and right-time automated actions.
Right-Time Data Integration
Right-time data integration supports the event-driven integration, propagation and migration of business transaction data and master data. One of the primary objectives of this style of real-time processing is to create a consistent and integrated store of business transaction and master data. The integrated store of data may be maintained synchronously or asynchronously with the source data, which in most organizations is dispersed across multiple systems. A synchronous data store will contain real-time current data, whereas an asynchronous store will contain low-latency data.
Right-Time Data Reporting
Right-time data reporting provides business users with on-demand access to business transaction and master data. The on-demand data may reside in current business transaction files and databases, or may have been extracted and integrated into a low-latency data store. The benefit of accessing current data is that it is real-time data. The advantage of accessing data that has been extracted into a low-latency store is that it is integrated and consistent. The downside of a low-latency store approach, of course, is that the data is not current - the data may be a few seconds or even several hours out of date, depending on the techniques used to maintain the store. For many applications, however, consistency is more important than timeliness.
Right-Time Performance Management
Right-time performance management enables business users to monitor and optimize the business performance of intraday business operations. It is one of three types of business performance management (BPM) - the other two types being strategic performance management and tactical performance management. The key attribute of right-time performance management (which is sometimes known as operational performance management) is that it is event-driven. As events occur in business transaction systems, they are sent to operational BPM applications for processing. These applications analyze the events and produce metrics about business performance in real time or at predefined intervals. Operational BPM applications can also be built using a data store created by right-time data integration.
Right-Time Automated Actions
Right-time automated actions improve the speed of business decision making and increase business productivity by automating the decision-making process. This is achieved by encapsulating business user expertise in a set of business rules that are embedded in a rules-driven workflow engine. As BI is produced (by a right-time performance management application, for example), it is passed to the rules engine for evaluation against associated business rules. These rules determine what action needs to be taken based on the results of the evaluation.
Technology Advances
Now that we have defined the four main categories of right-time BI processing, we are in a position to review the technology advances that have taken place over the past year in supporting them. The technologies we need to consider are BAM (business activity monitoring), BRE (business rules engines), EAI (enterprise application integration), EII (enterprise information integration) and ETL (extract, transform and load).
For right-time data integration, EAI and/or ETL can be used (see Figure 1). Although EAI technology was intended initially for application-to-application integration, it has also been used extensively for data integration and can, therefore, be used for building low-latency and consolidated master data stores. EAI technology is typically a component of an application integration platform, and leading vendors include BEA, IBM, Microsoft, Oracle and TIBCO.
Figure 1: Right-Time Data Integration
The issue with some application integration platform products is scalability and transformation power, and this is why there are an increasing number of relationships between application integration and ETL vendors. The ETL products provide scalability and transformation power, while the application integration products provide an event-driven front end. Application integration products also offer a wide range of adapters for accessing back- and front-office application packages.
Over the past year, the ETL vendors have been enhancing their support for event-driven and Web services-based data sources. Leader charts from analyst companies such as Gartner, IDC and META Group show the key ETL vendors to be Ascential, Business Objects, DataMirror, IBM, Informatica, Microsoft, Oracle, Pervasive/Data Junction and SAS. The inclusion of Pervasive is important to note - its recently acquired Data Junction product set has enjoyed significant success in the small and medium business marketplace as a low-cost but powerful ETL capability. Many ETL vendors now call their products data integration platforms to demonstrate their move toward supplying a more generalized data integration environment for both batch and event-driven processing.
Be the first to comment on this post using the section below.