What is an operational data platform? - IntelliStream
There’s a category of software most people outside heavy industry have never heard of: the operational data platform. If your work doesn’t involve turbines, pipelines, or substations, you can be forgiven for never encountering it. But it’s worth understanding, because it’s the thing that decides whether an organization can actually trust its own numbers.
An operational data platform is one shared model of everything an organization runs; its assets, the processes they drive, the business context that frames them, and every signal and event they produce, with lineage on every derived value.
In plain terms, when a report shows a number, you can track it all the way back to its raw source, complete with data quality flags that show exactly how reliable that data is. When something goes wrong, you investigate by following the relationships across assets, processes, and events, instead of joining five systems by hand.
Large enterprise vendors have sold one for a decade. They sell it well. They also sell it for seven figures, with eighteen month implementation timelines, proprietary storage you can’t easily leave, and vendor relationships you can’t unwind.
IntelliStream DataHub is the same category, executed differently.
It’s open source, under AGPL-3.0. You can read the code, audit it for compliance, run it air gapped, and fork it if we ever disappear. It’s built on standard components like PostgreSQL, Apache Pulsar, ClickHouse and Neo4j, and no proprietary storage that you’ll never be able to escape.
You can deploy it in your own cloud, on-prem, air-gapped, or on our own multi petabyte infrastructure in Stavanger, Norway, owned and operated by us. And it’s routinely 10× below what the enterprise incumbents charge. Not a discount, but a different cost structure.
In practice, teams use it for many practical things, some of them:
- Audit-ready KPI reporting, where every figure traces back to the raw signals it was computed from.
- Operational investigations, where you follow the relationships across assets and events instead of joining five systems by hand.
- Onboarding new data sources, where you drop them into the model once and every downstream report inherits the context.
- Enhance AI, where models train and run on data that already carries its full context and lineage, so a prediction isn't a guess pulled from a stripped-down table.
- Operation Forensis, where after an incident you reconstruct exactly what happened by walking the timeline of events across the assets and processes involved.
- Predictive maintenance, where live signals from equipment are scored against their normal operating envelope, so you act on a bearing or a seal before it fails.
- Forecasting, where the same modeled history that powers your reports feeds production, demand, and capacity projections; one trusted source for what happened and what's likely next.
- Build new applications, where developers build against one consistent model with lineage and quality flags already in place.
- Risk management, where you trace exposure back through the assets and signals that drive it, with quality flags on every input.
- Team Collaboration, where everyone works from the same shared model and the same definitions, so the argument stops being "whose number is right" and starts being "what do we do about it."
We're a four person team building IntelliStream in the open, because we believe this category deserves an open answer, and because the platform under your data should never be a black box.
The positioning, in one line: "Intellistream is the industrial data platform you’d build yourself, if you had the time.”
You don’t have to take our word for any of it. Just run git clone && docker compose up and you’ll have a live instance packed with sample data to play with in about ten minutes. Or, if you’d rather skip the setup, just get in touch. We’d love to hang out and help you get started!