The UFIS architecture glossary

UFIS both reads from other systems and writes approved changes back into them.

The operating policies that decide what counts as normal, what triggers action, and what needs approval.

How much one individual signal moves the overall picture of how well things are aligned.

How likely it is that this particular signal is actually true, as distinct from how much the source is trusted generally.

When two trusted sources disagree, UFIS keeps both, flags the disagreement, and passes it forward rather than quietly picking one.

A prebuilt adapter for a specific system or feed, so connecting it does not require a bespoke project each time.

The idea that the next big constraint is not more data or more compute, but getting systems to act together.

The information exists somewhere, but nobody can see how the pieces affect each other — so decisions are made on a fraction of what matters.

Software that runs above existing systems instead of replacing them, so each system keeps its job while UFIS makes them work together.

Understanding how something in one industry or system affects something in a completely different one — weather affecting freight, freight affecting pricing.

A detailed virtual model of one known environment — a plant, a mine, a building.

Working out that 'Pump 14', 'PMP-014' and 'pump fourteen' in three different systems are the same physical thing.

Separate organisations can share what they have learned without handing over their raw data or losing independence.

Rules are applied the moment data arrives, not after the fact.

A label on every signal saying how sensitive it is and who is allowed to use it for what.

Nothing acts on its own authority — recommendations and automated actions are checked against policy and approval rules first.

Because UFIS increases how much cloud infrastructure is used, the major cloud providers benefit from it rather than compete with it.

The doorway into UFIS: the connectors and agents that pick up information from every system, sensor, file and external feed.

Who is allowed to see what, who can approve what, and proof of what happened.

The plumbing and traffic control: moving signals around and pushing approved actions back out to other systems.

Where everything is kept, in whichever store suits the shape of the data and the speed it is needed at.

Where raw signals become meaningful: matched to the right asset, site or customer, and scored for quality.

The models that forecast what happens next, spot the abnormal, and work out the better option.

The apps built on top for a specific job — a maintenance planner, a grid operator, a logistics controller.

What people actually see and do: dashboards, alerts, plain-language answers, and actions that fire in other systems.

A measure of how well the whole operation is working together, not just how each part is performing on its own.

The platform remembers what was decided, what was done and how it turned out — and uses that next time.

Turning a conclusion into coordinated action across several systems, rather than just showing someone a chart.

What actually happened after an action is fed back in as a new signal, so the system learns from it.

A patent application has been filed. It has not been granted, and pending status does not create enforceable rights.

The heavy original file — a satellite image, a long document — stays where it is; the signal object keeps a reference to it.

Spotting that things are about to go out of alignment before the problem actually shows up.

Sensitive detail can be turned into a usable score without exposing the underlying personal or classified data.

The record of where a piece of information came from and how it was handled along the way.

The map of what affects what — built and maintained automatically, including links nobody defined in advance.

Whether UFIS interpreted the signal correctly — separate from whether the signal itself was right.

Any measurable change worth noticing — a sensor reading, a forecast, a price move, a human approval.

Signals do not live forever — they are created, checked, connected to other signals, acted on, and eventually retired.

How reliable the source of a signal has proven to be over time.

A coordination layer that sits above the systems a business already runs and helps them understand each other.

A single common format that every event is translated into, so a rainfall forecast and a machine vibration reading can be compared side by side.

The stage that makes very different signals comparable — aligning time, place and units — without pretending they are the same thing.

The standard fields every signal carries: where it came from, when and where it happened, how much it can be trusted, and what it is connected to.