Aggregate Analytics with Digital Twins

In addition to providing deep introspection during message processing, the dynamic state information held in digital twins creates the basis for real-time, aggregate analytics. Instead of moving this state information to an offline store for batch analysis, which can require minutes or hours, the cloud service’s execution platform can perform continuous, data-parallel calculations (e.g., MapReduce) on the state of all digital twin instances every few seconds to identify aggregate trends and maximize situational awareness. The results of this aggregate analysis allow managers to immediately sift through large volumes of data and quickly focus on areas of concern. This also provides additional, timely feedback for use message-processing.

For example, in a streaming analytics use case, a rental car application can perform continuous, aggregate analysis across all real-time digital twin instances (as illustrated below) to quickly pinpoint emerging trends, such as highway blockages due to accidents or floods, that may require immediate attention or feedback to the vehicles. The ability to triage the state of an entire fleet within a few seconds enables monitoring personnel to immediately focus attention on the most important problems and not overlook critical issues due to the sheer volume of incoming telemetry.