|
|
| Home - Industry Article - Feb 07 Issue |
How can CEP Transform Business Process Automation |
By David Cameron, Vice President of Product Integration, AptSoft Corporation
Technology has long been used to automate business processes. However, with the advent of a new technique, event-driven applications using Complex Event Processing (CEP), developed at Stanford University, the requirements that can be economically addressed have increased substantially. For the first time, CEP allows technology to address a new class of business challenge central to today’s evolving business climate. CEP requires specific functionality and a novel way of designing application solutions, but organizations have successfully implemented it and subsequently realized its transformative qualities.
Introduction
There are many reasons to develop an application that automates a business process. First, technology can reduce the time and effort required by people performing repetitive tasks, a great productivity saver. Second, technology can reduce errors and by extension, reduce costs. Finally, technology can ensure a process runs ONLY as intended, either for internal compliance reasons or to conform to external legal or regulatory frameworks.
Given the plethora of tools on the market that facilitates the development and deployment of technology, there is no shortage of business requirements to be addressed. However, a class of business challenge has emerged with special requirements that, using traditional tools and approaches, necessitates a higher level of investment in development and maintenance than the perceived benefits, thereby remaining uneconomical to develop.
This class of Complex Event Processing, or CEP, application requires an event-driven approach using techniques compatible with, but currently not native to, traditional tools and approaches.
Event-driven Applications and Complex Event Processing
While event-driven applications are not new, the advent of CEP has made them more relevant than ever before.
CEP was first developed at Stanford University by Professor David Luckham in the late 1990s under a DARPA grant. Initially, the work was focused on developing a system that could detect patterns of activity on IT networks that indicated attempts at intrusion, and upon detection, trigger automated countermeasures. And while CEP includes functionality that seems similar to traditional tools like Business Activity Monitoring (BAM), Business Rules Engines (BRE), or Business Process Management (BPM), when used by a platform for designing and executing event-driven applications, CEP supports the development of a whole new class of applications.
In fact, CEP has made these applications so much more practical, and thus improved the cost-benefit calculus to such a degree, that the approach can often reduce the up-front and the on-going cost by a factor of 10, making it worth a second look.
The Characteristics of CEP Applications
Identifying the use for this approach is relatively easy once you know what to look for. Event-driven applications using CEP share one or more of the characteristics listed below in figure 1.
| The application flow is externally driven by unpredictable and non-linear activity, such as the automation of shipping and logistics activity where many components (equipment, cargo, labor, etc.) need to come together in real time in order for a task to be completed |
| The business logic changes quite frequently, such as in suspicious activity patterns used in fraud and compliance applications |
| The processing needs to ‘adapt’ to external circumstances in a ‘closed-loop’ environment, such as scheduling applications which must respond to weather, equipment, and resource availability issues ‘on-the-fly’ |
Figure 1: Application Characteristics Best Suited to CEP
Applications with these characteristics tend to defy conventional Information Technology development toolsets due to their complexity. Often, though not exclusively, areas that benefit greatly from an event-driven approach can be found in the areas of fraud detection, policy compliance, and customer acquisition and cross-sell.
Two Use Cases
Organizations have adopted two approaches to CEP. The first approach is project-centric: identify a project, usually one involving the above characteristics, and evaluate if it’s fit for CEP. The second approach is architecture-centric: build an architecture that transitions from a data-driven approach to building applications (moving data en masse from application to application and relying on people to Interpret and respond to that data) to an event-driven one (capture and expose key business and system activity as ‘events’ with contextual information like time and sequence, and process them in an event engine with functionality such as event decisioning, process flow, and monitoring). Both approaches have been successfully utilized.
In the project-centric case, Circle Company Associates, Inc., a provider of loyalty management solutions, needed a way to quickly address each client’s changing business requirements. The company’s value proposition is built around the real-time use of information to drive business processes. Its suite of concierge, personal assistance, and incentive and recognition services engage the customers and employees of Circles’ clients in a personal and memorable way to generate value from increased loyalty.
Through the use of key operational systems, Circles forges an ongoing dialogue, providing client organizations with an opportunity to learn more about their customer’s and employee’s needs and preferences. Circles’ personal assistants then use these insights to offer personalized time-saving services, making its clients indispensable. In one case, this involved a three-step process to:
- Verify membership
- Provide services
- Delivery closed-loop feedback with real-time disposition data sent back to Circle’s client
Because customer requests are event-driven, Circles implemented CEP to pull, package and send the data to the client’s web service for each step – eliminating the need for Circles’ service professionals to re-enter data into a separate system – and providing the real-time feedback required by the client.
In the architecture-centric case, ISO New England (ISO-NE), a US regional electricity transmission organization, created an architecture that publishes system events directly from the applications at each stage of the critical business processes involved in allowing participants in ISO-NE’s energy futures market to hedge themselves against the next day’s electricity price fluctuations. These events are then managed into a larger business process and monitored for complex patterns that may indicate issues or exceptions. A CEP engine is used to synthesize event patterns and analyze streams of events for the purposes of reporting, identifying exceptions, and generating alerts.
On January 18, 2006, a fast-moving sequence of events eluded human detection and shut down ISO-NE’s trading in the final minutes before the deadline. The use of complex event processing to rapidly intercept and respond to this event pattern will prevent this scenario from happening in the future.
The adoption of CEP is still in its early stages, but as these case studies illustrate, the approach has real value in addressing a new class of very important business problems. As more organizations realize the unique benefits of the event-driven approach in today’s hyper-competitive market, growth will accelerate.
|
|
|
