Skip to content
Proxus 2.0.1 Licence and sign-in fixes after the 2.0 upgrade Release notes
Try free Book a demo

Platform · Edge

Decide at the line, not three systems later.

The edge runtime collects, evaluates and buffers next to the equipment, and keeps doing so when the link to the centre drops. Gateways keep no configuration of their own: it comes from the hub, so a failed unit is replaced rather than rebuilt.

Book a demo What happens when the network drops
Screen recording: a pipeline that runs at the line, in plain words with what it would do now, then as a canvas, then the gateway it runs on.
Runs
One gateway process plus a messaging container next to it that holds the local buffer, installed by one script on a plant server or industrial PC. No database at the edge
Evaluates
Rules, alarms, statistical anomaly detection and your own functions
Keeps working
Collection, rules and alarms continue when the link to the hub drops. A gateway that restarts while the hub is unreachable waits for it before collecting
Buffers
48 hours of data by default, up to 30 days per gateway, replayed when the link returns

A decision at the line

The gateway decides, then tells the hub.

A pipeline runs on the gateway next to the machine, so the alarm does not wait for a round trip to the centre.

  1. Reading MFB_LYS_TUNNEL_OVEN · Zone1Temp 182 °C Read by the gateway at the line
  2. Pipeline Above 180 °C, until back to 175 °C Runs on the gateway
  3. Decision Warning alarm opened At the line, in the same moment
  4. Hub Alarm and values arrive Held at the gateway while the link is down

In the product

Try it, write it, see where it runs.

See what it would do, before it runs

Beside every pipeline, what it would do with each device’s latest values right now. Nothing is sent, written or opened.

  • A sandbox runs a draft on samples, with external effects blocked
  • Each step says what it would change before you publish

Shape the data at the source

Filter, keep or rename fields, convert types and calculate new values in a pipeline on the gateway, before anything leaves the line.

  • Each target then sends every reading or only changes
  • Runs on the gateway, next to the machine
A deployed C# function in the editor, with its Build and Deploy actions.

Your own code, at the line

When a ready-made step is not enough, write a C# function. The hub compiles it and sends it to the gateways you choose.

  • Compile errors with their line and column
  • Build and deploy are separate permissions per role
The SYSTEM gateway’s overview with the devices behind it. Enlarge
The SYSTEM gateway’s overview with the devices behind it.

Every gateway at a glance

Each gateway shows whether it is online, how many of its devices send data, its share of the licensed rate, and its CPU and memory.

  • Devices, functions, rules and pipelines it runs, one tab each
  • Buffer retention set per gateway, 1 to 720 hours

Real Proxus screens; the machine data shown is simulated.

Want to plan this for your sites? Plan your gateways

When the link drops

An outage delays data. It does not erase it.

The gateway keeps reading and holds each value with its own time, 48 hours by default and up to 30 days. When the link returns, the values reach the hub in order and the history has no gap.

Step through an outage

Online: every value reaches the history as it is read.

Technical details Problems it solves, every capability and the full protocol table

What it solves

What teams run into without it.

  • Latency where it matters

    A rule that has to wait for a round trip to a data centre is not a rule an operator can act on.

  • Outages become holes

    When collection lives centrally, every network incident writes itself permanently into your history.

  • Bandwidth spent on noise

    Sending every raw sample to the centre costs link capacity and storage for values nobody reads.

  • Every site becomes a pet

    When configuration lives on the box, a failed PC means travelling to site, reinstalling, restoring a backup that may not exist and reconfiguring drivers by hand.

What you get

Built in, not bolted on.

  • Rule engine

    Expression rules (thresholds and conditions) and anomaly rules, evaluated as each value arrives, raising alarms or triggering downstream actions. Before you save, a rule is replayed over the last 24 hours of stored data, so you see how often it would have fired.

  • Anomaly detection

    Z-score, IQR, moving average, rate of change, drift, flatline and gap detection, running live at the edge as values arrive. The trend workspace shows the stored alert moments.

  • Custom functions

    Write C# functions against the Proxus SDK when a standard block is not enough.

  • Store and forward

    A local buffer keeps 48 hours of data by default (set it per gateway, up to 30 days) and replays it when the link returns, so an outage delays data instead of destroying it.

  • Transform before transport

    Filter, calculate, convert types and set units in a pipeline on the gateway, and choose per target whether every reading goes or only the values that change.

  • Fleet of gateways

    Deploy additional gateways per site or per area, each with its own identity, all managed from the centre.

  • Configuration from the hub

    A new gateway authenticates, receives its devices, rules, functions and integrations, and starts working.

  • Replace, do not repair

    Hardware fails eventually. Install the replacement with a new command from the hub and it resumes with the same devices, rules and functions.

What this enables

What the capabilities add up to, in practice.

  • Install the site with one script and update it as a unit
  • Replace a failed gateway without rebuilding its configuration
  • Keep collecting and evaluating when the network or the centre is down
  • Act on a rule within the process cycle instead of after a round trip
  • Send curated data upstream rather than every raw sample
  • Extend the pipeline with your own logic without leaving the platform

When the link drops

An outage delays data. It does not erase it.

  1. The link drops

    A WAN fault, a switch reboot or maintenance on the central server: the gateway can no longer reach the hub.

  2. The line keeps working

    Drivers keep reading, rules keep evaluating and alarms are queued. Every value is written to the gateway’s local buffer.

  3. The link returns

    The buffer replays to the hub and queued alarms are delivered. Nobody exports or re-imports anything.

  4. History is filled in

    Within the buffer retention, the historian fills in the outage period from the buffered values, so next month’s report has no gap to explain.

Buffer retention is set per gateway in the hub: 48 hours by default, up to 30 days, as long as the gateway’s disk can hold that much data at its rate. An outage longer than the retention loses its oldest values first. The internet is not part of this path: with the hub on your own network, an internet outage does not interrupt collection at all.

How it is built

Proxus is written as one system, not assembled from other products: one gateway process at the line, a messaging container beside it for the buffer, and no database at the edge.

Read the architecture

Proven in production

Running at the edge, including where bandwidth is scarce.

All deployments

Questions

What teams ask before the first call.

Ask something else
What runs at the edge and what runs centrally?

Collection, rules, anomaly detection, custom functions and buffering run on the edge gateway next to the equipment. The namespace, historian, dashboards and central administration run on the hub. The same product, two roles.

How much hardware does a gateway need?

It runs as one gateway process plus a messaging container that holds the local buffer, installed by one script on a plant server, an industrial PC or an edge device. Sizing depends on how many signals and what interval you collect at, which is one of the things a scoping call establishes.

Can we write our own logic?

Yes. C# functions written against the Proxus SDK run in the pipeline alongside standard blocks.

Continue

Two numbers set your price: the gateways you run and the data rate they carry. See pricing

Plan your gateways

How many gateways, and what is the network like?

Tell us how many sites and gateways you expect and what your network allows. An engineer replies within one working day.

Your network

We use your email to answer this check. Marketing emails only if you tick the box; no resale. Privacy notice

Plan gateways Live demo