Device count multiplied by reporting interval is what actually sizes a telematics server and a SIM plan. This is that arithmetic, made visible.
Change any input and the result updates. Nothing is emailed anywhere and nothing is stored.
Record size covers a typical position plus status flags; sensor-rich records and CAN data run larger. Database figures apply a 2× multiplier over raw payload for indexes, timestamps and row overhead, which is the usual shape on a relational store.
Sales pages talk about CPU cores. The number that actually decides your server is the one this calculator produces: how many rows per second arrive, and how many are retained. A thousand devices on a 30-second interval produce close to 2.9 million records a day. The same thousand on a five-minute interval produce under 300,000 — a tenfold difference in write load, storage and query cost from one setting.
This is why we ask for device models, reporting intervals and retention period before quoting hosting, and why a provider who quotes without asking is guessing.
Shorter is not better, it is more expensive. Thirty seconds suits urban delivery where the route detail matters and a missed turn is a customer complaint. Sixty to one hundred and twenty seconds is ample for long-haul, where the vehicle travels in a straight line for hours. Battery-powered asset trackers reporting every few hours are behaving correctly, not failing.
Most platforms also report on event as well as on interval — ignition, harsh braking, geofence crossing, panic. Those are additive to the figures here, and on a fleet with aggressive alerting they are not a rounding error.
The per-device monthly figure above is raw payload. Real cellular consumption runs higher because of protocol overhead, keepalives, reconnections and re-sent buffered records after coverage gaps — expect meaningfully more than the raw number, and size a plan with headroom rather than exactly. Even so, a standard position tracker is a modest data consumer. Video telematics is the exception, and it changes the arithmetic entirely.
Storage is not driven by how much data arrives but by how long you keep it. Doubling retention doubles the database. Set the period from your actual dispute, insurance and contractual window rather than from a default, because an event you cannot produce six months later is an event you cannot act on — and because paying to store five years of data nobody queries is a recurring cost with no return.
Send us the numbers this produced and we will size a server against them properly. Request a hosting quote or message us on WhatsApp — device models and reporting intervals are enough to start.
For a standard position tracker, modest — typically single-digit megabytes at common reporting intervals. Real consumption runs above the raw payload figure because of protocol overhead, keepalives, reconnections and re-sent buffered records, so size a plan with headroom. Video telematics is a different category entirely.
Thirty seconds for urban delivery where route detail matters. Sixty to one hundred and twenty seconds for long-haul, where the vehicle travels predictably for long stretches. Hours for battery-powered assets. Shorter is not better; it is more expensive in data, storage and query load, and the extra detail is frequently never looked at.
Indexes, timestamps and per-row overhead. A position record is not stored as bare payload — it carries its own identifiers and the indexes that make it queryable, and those are what make historical reports fast. A 2× multiplier is the usual shape on a relational store; heavily indexed schemas run higher.
Yes. Ignition, harsh braking, geofence crossings and alarms report in addition to the interval, so a fleet with aggressive alerting generates meaningfully more than the interval arithmetic alone suggests. Treat the figures here as the floor.
Longer than your dispute, insurance and contractual window, and no longer. Retention is a straight multiplier on storage cost, so keeping five years of data nobody queries is a recurring expense with no return — while being unable to produce a six-month-old event makes the whole system less useful at exactly the moment it matters.
Send us what you put in. We will tell you where the assumptions are optimistic, including when that argues against buying from us.
Discuss Your Numbers →