Scaleout Software has announced in-memory digital twins to receive live operational telemetry and make it available to generative AI models
The company says retrieval-augmented generation (RAG) relies on traditional data sources, such as databases, that cannot provide live telemetry at scale, or on streaming platforms, which lack descriptive information about the system in their telemetry database. As a result, AI lacks the continuously updated, contextualized view of a live system needed to effectively monitor and analyze it. Scaleout’s digital twin SW removes this obstruction.
Scaleout says v7 of its software uses an object-oriented, digital twin model hosted in memory to track all system components in real time. Digital twins can incorporate metadata that describe the overall model, individual digital twin objects, and their properties, including details about each property’s units and expected min/max values. This gives AI RAG instant access to live data by interrogating component properties within the digital twin, meaning it can access system structure context, and expected behavior; its state and structure.
Dr. William Bain, CEO and founder of ScaleOut Software, said: “Operational managers are now using AI models to monitor and analyze large systems, like transportation and logistics systems. However, these models face major bottlenecks ingesting live data and capturing context about a live system. Digital twins have the speed and richness to meet this challenge and unlock AI’s potential.”
ScaleOut hosts the in-memory digital twin on a cluster of servers and automatically scales performance as needed to match the size and complexity of a live system. This ensures that data access remains fast for very large digital twins with thousands or even millions of digital twin objects.
The digital twin uses in-memory objects to track individual components within the data center, and it periodically receives telemetry from these components. Each digital twin object also maintains contextual information, such as model numbers, specifications, and connections to other components. Starting with version 7, each digital twin object also hosts metadata that explains the role of each component, its expected behavior, and the structure of the component hierarchy.
Version 7 now supports access to digital twin data using the industry-standard Model Context Protocol (MCP) with authentication for Claude Desktop in this release. Support for other AI models will follow in upcoming releases.
V7 of Scaleout’s software also includes anew integrated user interface which unifies management of ScaleOut Digital Twins within the ScaleOut Product Suite. Developers can manage digital twin models alongside other platform features, including ScaleOut Active Caching for managing fast-changing application data.
Developers can also deploy data-parallel methods across entire collections of in-memory objects to implement fast, scalable analysis of cached data.
A blog provides more information about the v7 software.
Bootnote 1
Scaleout Software was founded in Bellevue, WA, in 2003 by Dr. William Bain and is privately anfd(limited) angel funded. Bain founded and ran Valence Computing from 1992 to 1996, then Valence Research from 1997 to 1998 with web load-balancing software, when Microsoft acquired it. MIcrosoft incorporated it as Network Load Balancing in Windows Server. He spent almost 4 years at Microsoft as a SW development engineer and program manager before starting up Scaleout Software and funding it privately.
Scaleout provides scalable, highly available distributed caching, in-memory computing, and digital twin technologies and has an in-memory computing platform which stories, updates and analyzes changing, live data.
The ScaleOut Product Suite focuses on scalable, highly available in-memory data grids, distributed caching, data-parallel analytics, and digital twins. Core products include:
-
ScaleOut StateServer — In-memory data grid and distributed cache for Linux and Windows, with high availability, automated cluster management, and “Active Caching” (running application code and custom data structures in the grid).
-
ScaleOut StateServer Pro — Adds data-parallel computing so analysis happens where the data lives.
-
ScaleOut GeoServer — WAN replication and global data integration across data centers for disaster recovery.
-
ScaleOut Digital Twins — Stateful stream processing that models each data aggregation, simulation, and feedback. Used for IoT, telematics, logistics, healthcare, and similar workloads
-
ScaleOut In-Memory Database — Redis-compatible clustering with automatic management and consistency features, built on their grid architecture.
Products run natively on Windows and Linux (C-based server implementation), support Java, C#, C++, and REST APIs, and deploy on-premises or in AWS, Azure, and Google Cloud.
Customers use the software for session state and personalization in e-commerce, real-time reservations and tracking in travel/airlines, asset and issue tracking in telecom, IoT/fleet monitoring, and other operational systems that need low-latency access and analysis of live data. Recent versions have added generative AI and machine learning. Version 4 integrated OpenAI models and ML for anomaly detection on digital twins.
Bootnote 2
Data-parallel computing is a style of parallel processing in which the same operation is applied at the same time to many different pieces of data. Instead of splitting work into many different tasks (task parallelism), you split a large dataset into chunks and run one algorithm across all of those chunks concurrently. The program logic stays the same; only the data differs.
In ScaleOut’s products, particularly StateServer Pro, data-parallel computing means analysis runs inside the in-memory data grid – where the live objects already live – rather than pulling data out to a client or separate analytics cluster. The grid partitions the objects, runs the same analysis on each partition in parallel, and returns aggregated results quickly. This turns a distributed cache into a platform for operational intelligence on fast-changing data.
