February 19, 2025
Many of today’s complex systems, from transportation networks to power grids and smart cities, have thousands of components that need continuous real-time monitoring. These systems typically collect real-time data from IoT devices and other data sources, store it in databases and visualize trends. They rely on operations managers to look for problems and take remedial action when they crop up.
Here's an example of a telematics system that tracks telemetry from a fleet of trucks:
Dispatchers query a database to look for problems, like lost or erratic drivers or mechanical issues with engines. Offline big data analysis computes important trends, like on-time performance and fuel efficiency, and periodically visualizes the results.
The problem with this approach to real-time monitoring is that is relies heavily on the “human in the loop.” As systems to be managed grow in size and complexity, it becomes increasingly easy to miss emerging issues and challenging to respond quickly. Live data sits passively in databases until operations managers extract it and analyze it.
What’s needed is autonomous streaming analytics that intelligently analyzes incoming data and alerts managers when unexpected issues (or opportunities) arise. That is what real-time digital twins can do.
Real-time digital twins are memory-based software objects that hold information about individual data sources. They also contain application code that analyzes incoming messages from these data sources and alerts managers when they detect emerging issues. They use stored information as context to enhance their analysis, and they can incorporate machine-learning (ML) algorithms to identify subtle issues.
For example, if a real-time digital twin for a truck reports unexpected lateral accelerations, it can check whether the driver has been on duty too long or has an unsafe driving history. It can then alert the dispatcher within milliseconds to contact the driver.
Now multiply the power of a real-time digital twin by 100 thousand. Nationwide trucking fleets (or rental car fleets) may have these many vehicles on the road. It’s impossible for a small team of dispatchers to watch everything. But 100K real-time digital twins can easily track all vehicles and pinpoint important issues for the dispatchers. Their ability to provide continuous real-time analytics at scale is crucial for safe and efficient operations.
As another example, consider a multi-state power grid with thousands of transmission towers and power distribution nodes. By continuously monitoring all of these components, real-time digital twins can detect issues, like overheated transformers or power fluctuations, that might lead to service interruptions or fires. They can give operations personnel faster and more reliable indications of emerging problems than manually monitoring complex visual displays.
Real-time digital twins hold dynamic, contextual information about individual data sources. This information enables deeper introspection on the state of data sources than just examining messages, and it provides an up-to-the-second record of changes to each data source.
State information held in real-time digital twins offers a rich source of insights for operations managers and can boost situational awareness. By aggregating this data and visualizing the results, managers can spot and respond to important trends in a few seconds. They can assess how widespread problems are and direct their resources to handle the most pressing needs.
For example, when fleet dispatchers start seeing vehicle delays, they can use data aggregation and visualization to immediately determine where and how widespread the delays are and home in on the possible causes (highway construction, accidents, flooding, congestion, etc.) They can make informed decisions on how to respond. The ability to triage the state of an entire fleet within a few seconds enables personnel to immediately focus attention on the most important problems and not overlook critical issues due to the sheer volume of incoming telemetry.
Here's an example of real-time data aggregation reported every few seconds for a vehicle fleet showing the average delivery delay in minutes by region of the U.S. In this example, a snowstorm in the north central U.S. has moved east into the mid-Atlantic and New England states, causing substantial delays:
.png)
ScaleOut Software’s in-memory computing technology provides the processing power to perform continuous real-time monitoring at scale. ScaleOut Digital Twins™ is a scalable, highly available software platform for hosting millions of real-time digital twins in memory on a cluster of commodity servers. This platform can run either on premises or in the cloud. It delivers incoming messages to their corresponding digital twins and generates analytics results in a few milliseconds.
At the same time, ScaleOut Digital Twins performs continuous data aggregation and visualization every few seconds. It also enables the optional use of generative AI to continuously examine the results of data aggregation and alert managers when AI spots anomalies that may be of interest.
ScaleOut Digital Twins’ software architecture simplifies the work of application developers. They can use open-source APIs to build digital twin templates in C#, Java, or using business rules and optionally incorporating ML algorithms. Once deployed, these templates enable the platform to create real-time digital twins and deliver messages to them. The developer just focuses on writing the code required to analyze messages from a single data source, and the platform takes care of the rest.
We depend on large, complex systems to keep our modern society running smoothly. Disruptions can be costly and must be kept to a minimum. It’s amazing how many systems lack automated, real-time monitoring to assist operations managers. Real-time digital twins can help meet this need. By keeping key, contextual data in memory for fast access and organized by data source, they can analyze incoming messages in milliseconds and quickly identify problems. They can harness the power of machine learning, data aggregation, and generative AI to boost situational awareness for managers so that they can swiftly address problems and resolve them effectively.
Schedule a customized demo here.
Dr. William L. Bain is the founder and CEO of ScaleOut Software, which has been developing software products since 2003 designed to enhance operational intelligence within live systems using scalable, in-memory computing technology. Bill earned a Ph.D. in electrical engineering from Rice University. With over a 40-year career focused on parallel computing, he has contributed to advancements at Bell Labs Research, Intel, and Microsoft, and holds several patents in computer architecture and distributed computing.
