Marcus Collins

The adoption of an event-driven paradigm and the analysis of data in motion will enable enterprises to make accurate decisions more quickly. This greater speed means that enterprises are better able to see and react to changes in the business environment and therefore gain a competitive advantage over organizations that perform the more traditional retrospective analysis of historic data. Although the benefits are great, the complexity of implementing an event-driven infrastructure and processing means that the implementation of such an initiative is fraught with technical risk. IT organizations should be in the vanguard of this adoption and lead the drive to explore areas within the enterprise where real-time event processing could be of potential benefit.

Forward-thinking enterprises are realizing that the retrospective view of the business environment—a view that is supported by traditional business intelligence—is not sufficient to meet the demands of operating in a competitive environment where decisions must be made at “Internet speed.” Compounding this issue is the fact that instrumented business processes and the increasingly sensor-enabled physical world have caused a dramatic increase in the volume of business events that are received as digital feeds. Real-time data analysis on streaming data is, therefore, quickly emerging as an important technology for IT organizations to understand.

Historically rooted in capital markets, real-time data analysis is now expanding into financial-services fraud detection, intelligence gathering, telecommunications-network monitoring, supply chain, and logistics.

Enterprises should not assume that real-time analysis is only for high volume data feeds. Analyzing low-volume events can generate significant business value. Tracking inventory movements at a palate level and on an hourly basis may be all that is required to drive increased value in a supply chain. The technical risk and cost of implementing this low-volume event scenario will be less than that required to track each inventory item, using RFID tags, as it physically moves through a warehouse. Ensure that the event rate and processing complexity is appropriate for the business process and focus on the low-risk scenarios, especially at the initial stages of the implementation.

Whilst real-time analysis is best suited for “internet speed” decisions, retrospective analysis of business event data can generate significant business value. Analyzing business events in order to understand the drivers of business change is of significant value to enterprises. Retrospective analysis can also be used as a low-risk approach and precursor to the wider deployment of real-time event processing in the enterprise.

As with all new technology, adoption initiatives start with a proof-of-concept initiative. Focus on business units with well-understood business processes where the adoption of an event-driven paradigm will yield measurable business value. As the initiative progresses and the business benefits can be shown to have been realized, use the initiative as a poster child to articulate the value of event-driven processing to the rest of the enterprise.

The real value of adopting an event-driven paradigm will be realized through the involvement of the business domain experts, especially in the programming of the real-time event processing system with the business domain knowledge and rules. The success (i.e., the “order winner”) of the real-time event processing initiative will be measured not in terms of the event throughput rates or the availability of the real-time event processing infrastructure (business will see this as an “order qualifier”). Rather, it will be measured in terms of the speed and accuracy of decision making and in the responsiveness of the enterprise to changes in the business environment.

I’ll be exploring real- analysis of data in motion in more detail in a presentation at Catalyst Europe currently going on in Prague.

1 Comment »

Category: Uncategorized     Tags:

Share