Introducing Business Activity Monitoring

Typically, an organization’s processes span multiple systems, channels, applications, departments, and external partners. In this case, how do we monitor such processes? What is the current state of the organizational processes? What is the benchmark for poorly-performing processes and exceptional processes? Most of the time, organizations are unable to answer such questions, or only have a vague idea for various reasons. Either they are monitoring the process with a very limited scope, or the mechanisms for monitoring the process are not in place to allow such details to be available. We rarely find organizations with process owners having an end-to-end view of a process. The big picture of a process is not available to the decision makers on a real-time basis.

Also, we have seen that a BPM cycle involves more than just automating a business process. Although modeling and analysis of the process plays an important part before a process is executed, the benefit is further highlighted by using Business Rules technology for added agility of the enterprise. One important factor that closes the loop for BPM is the aspect of monitoring the process on a continuous basis to pinpoint bottlenecks, as the process is executing within a business, and acts as a feedback for potential process improvement exercise. The need to monitor an organization’s business processes, especially as part of a larger BPM initiatives, is gaining considerable acceptability and demand. Such monitoring is the primary job of BAM.

What is BAM?

BAM allows a business to monitor its business processes, and related business events being generated in real-time, and provides an assessment of business process health based on pre-defined KPIs. This allows greater operational visibility of the business to relevant process owners for assessment and decision-making via real-time information dashboards. BAM also allows users to take actions based on information available on the dashboards.

Typically, systems providing BAM capabilities use business events to capture information from varied sources such as ERP, workflow, BPM, legacy systems, external partners, and suppliers. These data sources provide the necessary business measures, which are evaluated by the BAM against set KPIs, and provide the information in a user-friendly dashboard for the users.

BPM, SOA, and BAM

BPM, SOA, and BAM can be used as independent, isolated technologies. However, their benefits are compounded for a business if used together in a complementary fashion. As we can see in the following reference architecture, BAM works along with the services and process components to capture event-related information from a business and IT perspective, for analysis and reporting purposes.

In this case, SOA enables an organization to have a robust and flexible IT infrastructure that can help it easily achieve its BAM goals by allowing events and data from different services to be available to BAM for decision-making and realtime analysis. In an SOA-based solution, the business events will be inputs for BAM that are provided from the services layer. The linkage is via the BPM route, or through an event-based integration layer provided by an ESB.

We can refer to this relationship between BAM and SOA as being Service Oriented Activity Monitoring (SOAM), as today’s organizational setup will provide this event information to various BAM services interfaces exposed by the business applications in an enterprise.

In case of BPM, the business process describes the key activities required to fulfill the specified business action and its associated KPIs. These actions are executed as transactions using an orchestration engine and the underlying service layer. These transaction occurrences result in multiple process events to be created for each step within a transaction. BAM’s primary focus is on capturing, analyzing, and reporting on the transactions and events created by the process running over the SOA platform.

In case of BPM, the business process describes the key activities required to fulfill the specified business action and its associated KPIs. These actions are executed as transactions using an orchestration engine and the underlying service layer. These transaction occurrences result in multiple process events to be created for each step within a transaction. BAM’s primary focus is on capturing, analyzing, and reporting on the transactions and events created by the process running over the SOA platform.

BAM usually looks at collecting information about a process based on the following attributes:

• Quantity or Volume of Transactions or Events: One of the primary areas covered by BAM is the volume of events generated by a process. This is not just an IT metric, but more of a business-related metric to help business stakeholders analyze information points such as the number of orders shipped in a day, the number of trades made during trading hours, the number of helpdesk tickets closed by a call centre, and so on. Usually, we will define these KPIs in a process definition, and use BAM to raise alerts to the portfolio manager if the process is exceeding those values, for example, “Send an alert as soon as the stock portfolio value decreases more than 3% “.
• Time Bound Events: In this case, the BAM concentrates on time-related metrics such as helpdesk ticket process cycle time for high priority issues, general process cycle times, supply-related waiting time, and so on. Again, based on certain thresholds, alerts can be sent out, or the reports can be viewed by the management in real-time using customized dashboards.
• Faults: These are situations where the process is not running well. This could be due to a hardware fault, or a process related issue such as deadlocks, or some other issue. BAM helps in these scenarios by helping to identify areas of problems, and providing important metrics with respect to frequencies of such errors and their potential damage on process performance, and other dimensions such as cost, schedules, and so on.
• User Defined Events and Conditions: Apart from the general dimensions of volume, time, and errors in a process, a business user might want to define KPIs around specific business issues that need analysis. For example, for compliance requirements, a bank might be required to keep track of all high-value transactions to prevent money laundering. During implementation, the business analysts can define this KPI in the process model, which will then be implemented mostly by a rules engine, and the events generated will be used by BAM to provide statistical reports and dashboards based on the frequency of these transactions, specific regions and user types involved in such transactions, and so on.

The real value of BAM, however, does not come only from analysis of individual events, and exceptions generated during process execution. BAM provides a mechanism to correlate aggregated process events to help with a cause and effect analysis, pattern matching, and so on, which provides immense value to today’s businesses.

Although not in the scope of this article, another area which is gaining a lot of attention in this area is Complex Event Processing or CEP. This can be a perfect vehicle for implementing BAM in an enterprise in order to solve complex business issues. CEP is based on a concept of analyzing a set of specific events from a range of possible events, and identifying patterns that could be meaningful for an organization. Among the many applications of CEP, one example we can use is that of ‘Algorithmic Trading’, where CEP can be used to analyze a huge amount of market data, assess favorable patterns for trading, and initiate trading in a market based on this. A lot of banks are using this technology for performing low-value trades, and assessing risk positions simultaneously. CEP then records this information as a ‘fingerprint’, and maintains a history to use when deciding whether to execute similar trades in the future. As it gathers more experience and intelligence, the BAM tool supporting CEP can start to refine its predictive capabilities, and conduct more efficient calculations.