Posted by: Arun Gupta
EAI, enterprise application integration, real time integration, time value of information
Every now and then, there’s a flurry of activity, questions and debate around real time information—on inventory, sales, production, process approvals, financial metrics, and so on. The passionate appeals by vendors makes one wonder whether the business is really inefficient or missing out on a large opportunity by not disseminating information to the managers and CXOs in real time. Add to this the new dimension of “complex event processing”, and the picture depicts a Jurassic era of information enablement.
Real time information availability has been business’ aspiration for a long time. IT enablement of the processes and operations in an enterprise expedited availability, but batch processing still did not provide the information as the event happened. As the data mining tools matured and models appeared for predictive modeling, gaps of the present became very evident. SOA Integration and middle layer technology solutions reduced the time gap. Mobile computing removed the physical presence limitation, as trickles of information could be provided on the handheld.
Now cast an eye across industries and various processes that are fed with, or create information. We will observe that today information flows with every step, decision, and event—irrespective of the sector, size or geography, the paradigm is uniform. People create information, people consume information, and people transform information. Managers, supervisors, CXOs, and even customers, seek control with real time information availability. Is it necessary to provide real time information to all the stakeholders? How does it change their behavior, decision or end outcome, if at all?
Take the case of retail. For a customer shopping in a store, price information on nearby stores in real-time is valuable, as it helps her get the best price for a product. To the retailer, a product sold is information, as it indicates that a customer has chosen a product from the shelf, and the stock count is down by one. Based on the supply chain’s agility, the retailer can use this information to plan for replenishment. The information can also be shared with a supplier who may use this snippet for planning next delivery and the impact on production schedule.
All this looks good in a one-one relationship, but when you multiply the dimensions, the complexity renders the simplistic scenario unviable as the optimization across the value chain has multiple constraints that operate on each decision point. Even when the collation and decision points can be automated, “complex events” have a way of making decision making a really difficult task requiring human intervention. In the above scenario, if the retailer received hourly information, will it materially impact the quality of decisions or process triggers (like a replenishment)?
The ground reality is that real time information does matter to an enterprise, but the rule cannot be applied for every byte of information. For a nuclear reactor, there is no other way. In case of a manufacturing plant, PLC data is, inventory data is not. Similarly for a financial institution, risk positions can build up quickly unless near real time monitoring exists, but a trial balance can wait for end of day. The application of technology for real time information is a good tool to be judiciously applied, and not get carried away by the use cases presented by the seller of the technology. If you are not doing it, get started, but ask the question at every stage. What changes with real time information?