Latency on the floor
A local question is answered locally rather than across a link.
Compute where the data is produced.
Edge is not a smaller cloud. It is a placement decision: what has to be answered locally, and what has to be true centrally.
A site that answers locally and reconciles honestly with the centre.
One data layer
Edge workloads need local reads, tolerance for an intermittent link, and a reconciliation story. PLOMID is one system with one storage contract, so what a topology has to do — what is answered locally, and what has to be true centrally — can be described rather than approximated.
A site that answers locally and reconciles honestly with the centre.
The site
Readings taken where the equipment is, on a link that drops.
Time series
The same shapes, held two ways. The difference is not the storage, it is where the agreement between them lives.
Every shape below is a surface of the layer, not a format to be converted into one. They are read and written together.
The models this workload names, the path a request takes through them, and the surfaces that speak to the layer.
The shape of the work itself.
The models this workload names.
One path from a request to the data it names.
How the workload reaches the layer.
Choose a stage to read what happens there, or a shape to see every stage that handles it.
Acquire
Readings taken where the equipment is
Time series
Readings taken where the equipment is
Local records needed for a local answer
What moved, when, and what conflicted
The durable record behind the site
Each one is a property of the system rather than a number somebody else has to measure for you.
A local question is answered locally rather than across a link.
The edge and the centre share a data model instead of diverging.
Synchronisation conflicts are recorded events, not silent differences.
Where a workload runs becomes a decision that can be stated, checked and changed.
Edge is where deployment matters most: local reads, a link that drops, and one durable record behind both. The fabric that places workloads across sites automatically is still being built.
Derived from the industry pages that reference this solution, so the two directions of the relationship always agree.