Analyzing information in many organizations is carried out retrospectively. This ability to simply analyze historical data is likely to provide valuable information, but far too late for it to be actually acted on in a meaningful way. What is needed in many cases is the ability to detect events as they happen, and relate them to historical trends to help with the decision-making process.
This capability is often referred to as complex event processing (CEP). Of course, ever since the beginning of message-oriented middleware, it has been possible for events to be published to message queues as they happen, and for other activities to subscribe to those queues, in order to receive messages that relate to the topic that the activity is interested in. An example of this could be in a financial services application that subscribes to a newsfeed about a particular market sector.
However, in other areas, the sequence of events, and the fact that something has not happened that was expected to, may be significant. An example is in the area of baggage handling and passenger services at an airport, where the fact that a customer has checked in for a flight, with one suitcase, but has not reached the departure gate by five minutes before the airplane is due to take off, could be a significant event. The options may be to call the passenger via the public address system, but alternatively it may be wise to make arrangements to remove the suitcase from the hold. The concept of the ‘non-event’ is valuable here, and could be highly relevant in other areas – it relies on the detection and understanding of ‘normal’ patterns and alerting someone so that a decision can be made.
A number of vendors provide offerings in this area. For example, BEA Systems is starting to use the fruits of its micro-service architecture, and has recently announced event-driven capability within its WebLogic range of products, including the ability to process non-linear events.
Progress, with its Apama product, also provides the capability to rapidly process a stream of events. An example of the application of Apama from the financial services area is a trading system that generates recommendations to carry out a trade. Behind this, the organization has a risk management system, which has the objective of ensuring that the company is not over-exposed to risk.
When the trading systems get messages to conduct a trade, the risk management system could also monitor those events to ensure, for instance, that the company never holds more than a given percentage of a particular actively-traded market. Before the automated trading system could exceed the relevant limit, the risk management system could send a message to stop the trade happening.
Tibco also offers CEP with its BusinessEvents product, which allows real-time aggregation of information from distributed systems, so that seemingly unrelated events can be correlated.
In many business areas, the need to react to events (and non-events) will be important, yet the challenge will be to gain a clear understanding of what is relevant and important, and what is just ‘noise.’ The combination of the ability to act in real-time, together with the in-depth analysis of complex underlying data, is where the goldmine will be found.
Source: OpinionWire by Butler Group (www.butlergroup.com)