A mobile machine works in a field, a tunnel, a warehouse: where the network is weak, intermittent and billed by volume. Collecting its data in real time without losing it, or blowing up the subscription cost, is an engineering problem in its own right.
The field’s network constraint
In the field, connectivity is never guaranteed: partial mobile coverage, dropouts, limited bandwidth billed by transmitted volume. A naive, continuous upload of every frame is both unrealistic and costly. As ALAMO Group points out, part of the value of a solution is that telemetric transfer happens over low bandwidth, which keeps the subscription cost down. The network constraint is not a detail: it shapes the entire architecture.
Decoding and buffering at the source
The answer lies in the intelligence embedded in the connectivity gateway. CAN frames are decoded locally, at the source, then buffered. If the network drops, the data is stored and re-sent once the link is restored, following a store-and-forward logic: nothing is lost. Only the useful information is transmitted, deduplicated and compacted, rather than the full raw stream. This is the principle of digital sobriety applied to collection: reducing the transmitted volume without impoverishing the usable information.
From stream to usable time series
On the server side, the messages fed back through a lightweight MQTT-type protocol feed time-series Big Data storage, able to absorb large volumes and allow real-time or deferred processing. It is this end-to-end chain, from onboard decoding to the time-series database, that then enables visualisation, detection of critical parameters and predictive maintenance. The robustness of the collection is invisible when everything works, but it is what determines the reliability of everything built on top of it.
Frequently asked questions
What happens when a connected machine loses the network?
The data is not lost. The connectivity gateway decodes and stores the measurements locally, then re-sends them once the link is restored (store-and-forward logic). This keeps a continuous history even under intermittent coverage, which is essential for mobile machines operating in poorly covered areas.
Why not simply upload everything continuously?
Because mobile bandwidth is limited and billed by volume: transmitting all raw frames would be unreliable and very costly at fleet scale. By decoding and filtering at the source, only the useful information is transmitted, which reduces subscription cost and footprint without degrading the analysis.
