Complex Event Processing
Business Activity Monitoring and Business Intelligence
By Harpal Kochar, Senior Principal Product Management, Oracle
In today’s highly competitive, service-oriented business climate, operational managers and executives demand visibility into the status of their business process networks as they relate to the business key performance indicators (KPIs) in real time. To stay competitive and provide better service operations, managers need ways to analyze emanating application events so they can compute higher-level, complex event aggregates and thresholds that affect the business KPIs, and so they can alert business users with related context to act upon.
For instance, organizations that run distributed global supply chains with Just-in-Time inventory practices must constantly monitor their inventory levels and correlate them to the bill of materials and replenishment requests they’ve sent to their suppliers and logistics partners. They must do this so they can continually make sure they have a balanced flow of parts and inventory throughout their entire supply chain. Similarly, telecommunications companies that provision new services and new customers must continually monitor their provisioning processes that touch hundreds of operational systems to make sure they have an up-to-the minute view of the status of outstanding customer service requests. This paper outlines the types of operational analytics and real-time visibility challenges that organizations face, and the ways that Business Activity Monitoring (BAM) and Business Intelligence (BI) tools can help solve these challenges.
BI refers to how IT technologies can leverage data into meaningful information that provides richer insights into how a business is doing today, as well as a sense of where it may be going tomorrow. BI solutions help you analyze large amounts of historical data, identify patterns, and understand trends that might affect the business. Analyzing sales and operations, understanding customer behavior, and identifying opportunities for revenue increases and costs savings are some of the areas where BI solutions can give business users interesting insights.
BAM is a real-time, event-driven extension of BI. The extension is subtle but powerful. BI products have the reputation of working only with historical data gathered from data warehouses, and being inflexible to change. BAM technologies, on the other hand, can provide real-time business analytics on information emanating from transactional data sources including Web services, message queues etc. BAM lets you correlate heterogeneous events and patterns for causalities, aggregates, and thresholds based on end-user-defined preferences and models. It can deliver the analyzed information and alerts in real time to business users when and where the information matters. It provides a platform that enables a structured and collaborative problem-resolution process that ultimately helps you optimize business processes.
Having a combination of BI and BAM technologies to get comprehensive business intelligence information from business processes is a competitive differentiator for most businesses. Staying ahead of the competition requires the ability to predict and respond to trends in the marketplace in real time, rather than react to isolated events or a set of events that have already happened. The fusion of the expertise offered by both BAM and BI technologies enables operational intelligence—in other words, the ability of a business to react to an event or process exception in real time.
The ideal use cases for operations intelligence—those that are interested in proactively catching business exceptions before or right after they occur—can’t rely solely on traditional BI systems that, in turn, run on a scheduled basis and not when exceptions occur. In some cases, it’s important to capture non-events and their impact on businesses. For example, in a call center business, if a valued customer’s call with an operator exceeds a threshold resolution time, a floor manager might be advised via a proactive alert to intervene and resolve the situation. Here, the call-ending event has not occurred and timely knowledge of this fact valuable to the business.
BAM can potentially use historic analyzed facts from a BI system as its measuring instruments to derive the described thresholds to the data it analyzes. BAM systems can integrate with data mining and OLAP tools to provide monitoring capabilities on top of the analyzed information to facilitate closed-loop analytics. For example, if a data-mining tool reports a list of suspected fraud accounts, this information can be fed into a BAM tool, and the system will make sure the accounts are frozen and the fraud detection agents are engaged. It can also make sure the accounts are fixed within a particular timeframe—and if that deadline isn’t reached, it can issue proactive alerts to interested business users.
Consider a supply-chain visibility and logistics example: A retail clothing business wants to enhance its operations intelligence infrastructure. Information analyzed by BI creates data for year-over-year consumer salary comparisons, market-basket analysis, and forecasts. This information can be passed as measuring facts to a BAM system. If the analyzed facts suggest that over the last six quarters, the average buying capacity of an area has dropped, BAM could trigger an action that intercepts a new order for purchasing certain luxury items in the order fulfillment system if the quantity is the same as previously purchased. Or, if BI-analyzed facts show that a particular item has disproportionate sales in two outlets over a period of time, BAM could issue a request to transfer more of those goods to one of the outlets.
In another case in the same retail business, BI is used to view transaction patterns that ultimately lead to inventory shortages and supplier delivery inefficiencies. Once the BI analyst reviews the reports and decides to take action, his response is typically a manual process. Optimizing the correction cycle requires sending out proactive alerts. Such alerts could be sent to inventory control managers when levels trend toward dangerously low levels, while an automated request for replenishments from alternate suppliers is also sent. Another way to optimize the correction process would be to let analysts take structured action, including starting a workflow process, and actually see the results impacting the effected KPI in real time.
BAM and BI together provide operational managers a versatile real-time monitoring and analytics toolset that can help them better analyze emanating operational information. These tools let an operational user work through large volumes of historic data to find the hidden truths through visualizations such as dashboards, and provide structured information patterns for KPIs, threshold violations, and event correlations with an emphasis on real-time alerting.
About the Author
Harpal Kochar is a senior principal product manager in Oracle's Server Technologies division. He has more than 10 years of experience developing software for distributed systems, enterprise software and integration technologies. Mr. Kochar holds a degree in computer science from the University of Mumbai, India.More by Harpal Kochar