Commercial fleets have quietly become data operations. Telematics, dashcam video, electronic proof of delivery, live routing, driver comms: the modern van or truck generates and depends on a constant stream of it, and the connectivity carrying that stream is usually the least engineered part of the vehicle. I have written separately about passenger transport and blue-light fleets; this piece is for the logistics and commercial operator whose vehicles earn revenue by the mile.
The problem is variability, not coverage
The threat to fleet data is rarely a total absence of signal; it is the unpredictability of the signal you have. A moving vehicle crosses shifting terrain of masts, bands and congestion, and single-modem routers handle it badly in two characteristic ways. The first is the reactive failover blackout: the router waits for a link to properly fail before recovering, and for thirty seconds or more the telemetry gap grows and the video upload stalls. The second is the sticky connection: the modem clings to a distant, weak mast because the session has not technically failed, while a stronger cell sits ignored. Both failure modes are invisible on a coverage map and constant in the field.
The definition of success for fleet connectivity is therefore stability, not headline speed. A connection that peaks at 100 Mbps and gaps for five seconds every ten minutes is worse for telematics and driver comms than a steady fraction of that. We engineer to the worst mile of the route, not the best.
Packet-level stability across carriers
SpeedFusion bonds multiple cellular modems, ideally across different operators, into one logical tunnel managed at packet level. Sessions live in the tunnel, not on any single carrier, so a degrading link changes the packet distribution rather than killing the stream. Bonding 4G with 5G gives the practical best of both: 5G throughput where it exists, 4G's superior rural reach where it does not. For the traffic that matters most in a commercial fleet, driver VoIP, live dashcam pulls after an incident, time-critical ePOD, WAN Smoothing duplicates packets across links so a handover wobble never surfaces in the application, and Hot Failover keeps every path warm rather than waiting to react.
Vehicle engineering realities
The vehicle install decides more than the configuration does. External antennas beat anything inside the cab or the box; a van body attenuates signal just as effectively as any Faraday cage. Power needs ignition sensing and wide voltage tolerance so the router shuts down gracefully rather than flattening a battery over a weekend or rebooting on crank. Vibration-rated connectors and automotive-grade certification (E-Mark on road platforms) are the difference between a three-month install and a seven-year one. And data policy is an engineering input for commercial fleets: per-application priorities so telematics always reports even when a dashcam is uploading, and per-SIM budgets so video traffic cannot quietly triple the connectivity bill across two hundred vehicles.
Visibility at fleet scale
InControl2 turns the fleet into one manageable estate: per-vehicle link health and location, per-SIM consumption with anomaly alerts, remote diagnosis of a unit parked three counties away, and firmware staged in depot-hours waves. The GPS-correlated link history is commercially useful in a way operators do not expect: it shows exactly where along your regular corridors the network underperforms, which turns recurring driver complaints into a solvable carrier-mix decision.
Rolling out across a fleet
Fleet deployments succeed on repeatability. We scope from the applications and the real route footprint, prove the design on a pilot subset, then template everything, configuration, SIM strategy, physical install spec, so vehicle two hundred behaves identically to vehicle one and any engineer can service any unit. We train the operator's own team on the estate, with our managed service as depth behind them.
The short version
Bond across carriers at packet level so no single network's bad mile becomes your outage. Install for the vehicle's electrical and RF reality. Write the traffic priorities and data budgets down. Manage the whole fleet from one screen with templates and staged rollouts. If your fleet's connectivity still depends on one SIM per vehicle and hope, get in touch for a scoping conversation.
Frequently asked questions
Is multi-modem hardware overkill for a delivery fleet?
It depends what the data is worth. If telemetry gaps, failed ePOD or unavailable dashcam footage carry real cost, the incremental hardware cost per vehicle is small against it. Role-based tiers work well: bonded units for high-value assets, simpler units where the stakes are lower.
Which carriers should a UK fleet combine?
The right mix depends on your actual corridors, which is what the pilot phase and GPS-correlated history establish. The principle is fixed: at least two operators with materially different coverage footprints, so no single network's dead zone is yours.
How do you stop dashcam video swamping the data budget?
Application-level priorities and per-SIM policies: telemetry and driver comms guaranteed first, bulk video scheduled or rate-limited, budgets and alerts per vehicle so anomalies surface in days.
Can this integrate with our existing telematics platform?
Yes; the network layer is transparent to the applications riding it, and InControl2 data (including GPS) can feed existing fleet software via API where a single combined view is wanted.
What does the driver experience change?
Ideally nothing, which is the point. The systems in the cab simply keep working through the handovers and dead zones that used to interrupt them.