Introduction

Use Digital Twins for Both Streaming Analytics and Simulation

ScaleOut Digital Twins™ is a new software platform for building digital twins of complex, live systems. Digital twins can run both real-time streaming analytics and simulation. You can use ScaleOut Digital Twins to build applications that simultaneously track telemetry from many thousands or even millions of data sources in real time to identify issues and opportunities and provide highly targeted feedback. You can also build time-driven simulations with large numbers of interacting entities. You can use simulations to validate streaming analytics prior to deployment in live systems or to build models that assist in making design choices and predictions.

ScaleOut Digital Twins adds unique value for applications which need to intelligently monitor or simulate large number of devices and data sources, track and analyze important state changes for each data source, and quickly identify emerging opportunities and threats using real-time aggregate analytics. Typical applications include IoT, fleet and asset tracking, disaster recovery, security, telecommunications, health-device tracking, financial services, crime prevention, and ecommerce.

ScaleOut Digital Twins adds important new capabilities for application developers that simplify the implementation of digital twins. ScaleOut’s unique combination of object-oriented design and in-memory computing enables deeper introspection and simulation modeling than previously possible. Object-oriented design simplifies development, and in-memory computing accelerates performance. When used for real-time analytics, it can respond to incoming telemetry from individual data sources within within 1-3 milliseconds and generate aggregate statistics every 5 seconds. ScaleOut’s patented, in-memory computing technology cost-effectively scales to track millions of data sources.

ScaleOut Digital Twins hosts digital twin applications developed using the ScaleOut Digital Twin Builder™ software toolkit or the ScaleOut Model Development Tool™. For real-time analytics, the service connects to data sources using popular event hubs, such as Microsoft Azure IoT Hub, Amazon AWS SQS, and Azure Kafka, and it also allows connections via an integrated REST messaging service.

To maximize situational awareness in both real-time analytics and simulation, the ScaleOut Digital Twins user interface (UI) enables the creation of continuous real-time aggregate analytics across thousands of digital twins with graphical updates every few seconds. Aggregate results also can be streamed to external visualization platforms, such as Microsoft Power BI. State information stored within a digital twin can be queried to report dynamic changes in the state of data sources to the UI.

An Example of Using a Digital Twin for Streaming Analytics

The following diagram illustrates the use of ScaleOut Digital Twins to track geographically distributed vehicles in a fleet using digital twins for a telematics application. The digital twin receives telemetry from individual trucks, analyze it, and generate alerts or feedback if necessary.The platform also continuously extracts and aggregates information from the digital twins in real time to boost situational awareness. Software objects, called digital twin objects and depicted as circles, track each vehicle in the fleet:

As the diagram illustrates, ScaleOut Digital Twins uses software objects to construct a digital twin of a physical system. This simplifies application design by allowing developers to focus on just building a set of software models, called digital twin object models that describe the state information and code needed to analyze telemetry from each type of physical data source. Developers focus on device behavior, and the in-memory computing platform takes care of the rest: creating objects for new data sources, delivering messages to the objects, running application code when a message arrives, scaling performance, and ensuring high availability.

Examples of Using a Digital Twin for Simulation

Digital twin objects also can be used to build a workload generator that runs in simulation to test real-time analytics. For comparison purposes, the following diagram shows a live system on the left in which digital twin objects track a fleet of trucks (with each object analyzing telemetry from a specific truck), and it shows on the right a simulation in which digital twin objects simulate the behavior of trucks and generate telemetry messages. These objects send messages to their corresponding real-time digital twins to test and evaluate real-time analytics. The simulated trucks can be parameterized to model various behaviors, such as a lost truck, mechanical issue, or an erratic driver. Once the real-time analytics has been thoroughly evaluated in simulation, it can then be deployed in a live system.

Digital twin objects can also model entities in a pure simulation used to assist in design and prediction. For example, an airline simulation might need to model passengers, aircraft, airport gates, and air traffic sectors to assess the impact of weather delays, outages such as ground stops, schedule changes, and more. Digital twin objects can maintain state information about the physical entities they represent, and they can run code at each time step in the simulation model’s execution to update their state. They also can exchange messages to model interactions. Here is a depiction of an airline tracking simulation:

Roadmap for this User Guide

The next topic provides an overview of ScaleOut’s digital twin model and explains its advantages for use in building applications for real-time analytics and simulation. This topic also describes advanced features, including machine learning for analyzing incoming messages and optional integration with the Azure Digital Twins service. The following topic describes the ScaleOut Digital Twin Builder software toolkit, which provides APIs and libraries for building real-time digital twins in C# and Java, and the topic ScaleOut Model Development Tool explains how to build real-time digital twins using an intuitive rules engine or with no-code machine learning using Microsoft’s ML.NET library. The Connectors topic explains how to connect data sources, such as IoT devices, to the ScaleOut Digital Twins service using multiple popular event hubs, such as Azure IoT Hub, or REST. It also explains how to optionally connect to the Azure Digital Twins service to integrate with that service. The UI Overview topic gives a tour of the service’s user interface, which lets users deploy digital twin models, deploy connectors for real-time analytics or start simulations, build widgets to graphically display real-time, aggregate statistics, examine digital twin instances, and manage the deployment.

Note

ScaleOut Digital Twins can be deployed on premises or in popular public clouds.

ScaleOut Digital Twins is available for production use. We encourage you to contact ScaleOut Software (sales@scaleoutsoftware.com) to learn more and to provide feedback. This will enable you to access the service’s UI and start building applications using this revolutionary new approach to real-time streaming analytics and simulation.