No datasheet states what a monitoring vehicle spends on data in a month. The omission is honest, because the number is not a property of the terminal. It is a property of the fleet’s habits: how many minutes anyone watches, how many alarms fire, how many investigations pull footage. The month gets assembled, from five components with known rates, then checked against the carrier portal’s measured figure. The assembly takes one sheet of paper written once. The check takes fifteen minutes a month for the life of the contract.
The published landscape confirms the gap. Equipment pages quote storage days and channel counts. Carrier pages quote pooled plans in gigabyte units, with marketed video-telematics pools sitting at gigabyte scale per vehicle. Neither side commits to a per-vehicle monthly figure, because neither side controls the habits that produce it. A fleet that wants its number builds it, on the per-activity rates the billing page of this series already prices.
The assembly answers three working questions, each one read off the same sheet. What plan size fits this fleet, before the first invoice arrives. Which vehicles are behaving abnormally, once invoices exist to compare against. When the configuration is wrong, where in the five components the waste sits. One model serves all three, read in different directions.
The components and their worked sums sit below, followed by the reconciliation against the portal, the outlier diagnosis and the plan-sizing rule. Every rate below is the series’ standing figure. Every total is stated with its quantity assumptions written beside it.
The month is assembled from five parts, then measured once.
Component one is the resident floor, the traffic a vehicle produces by existing on the platform with nobody watching and nothing happening. The heartbeat exchange runs in single-digit megabytes per month at common intervals, on the heartbeat page’s arithmetic. Position reports add points of tens of bytes at the tracking cycle, summing to tens of megabytes on busy reporting profiles. Protocol overhead, registrations and parameter reads ride along at similar scale. The floor lands under a tenth of a gigabyte for ordinary profiles, present on every vehicle every month, moving with configuration alone. Component two is supervisory viewing, the live sub stream watched from the centre. The rate is a few hundred kilobits per second per channel, the watch tier the codec and billing pages price. The driver of the total is minutes: a vehicle watched ten minutes a day at that tier accumulates a fraction of a gigabyte across the month, with the figure scaling linearly in watched minutes and in channels watched at once. Nobody watches the average vehicle for long, the reason this component stays small per vehicle. Control-room staffing decides its fleet-wide sum. Component three is alarm parcels, the evidence channel of the whole system. Each event packages stills, a clip and a record, in the megabytes per event on the consumption page’s figures. An ordinary vehicle fires a handful of confirmed events a month, putting the component around a tenth of a gigabyte. The component scales with the event rate, the trigger sensitivity and the pre-event window, the settings the storage trade-off page walks. Queued parcels from coverage holes arrive late and count in the same component, moved in time and unchanged in size. Component four is named pulls, stored footage retrieved across the link by case number. A pulled stretch crosses at recording-tier size, fractions of a gigabyte for minutes of full-quality footage. The component is zero on a quiet month and jumps when an investigation lands, which is why it is budgeted per case in the operations procedure and read separately in the portal. Component five is the exceptional item: a firmware update delivered over the air, a roaming day at a border, a misconfiguration streaming where nobody asked. The honest budget for the class is a small allowance plus the diagnosis habit below, because its defining property is irregularity. The five components close over everything a SIM carries. Any measured month decomposes into them, which is what makes the model an audit tool and a sizing tool with the same five lines.
The city van’s month assembles low on every component. The floor sits under a tenth of a gigabyte on an ordinary reporting profile. Supervisory viewing is occasional, a few minutes here and there, summing to one or two tenths. Alarms fire at city rates, a handful of confirmed events, another tenth or two. Pulls are rare, with the depot card reader covering deep investigations at no uplink cost, the lean the trade-off page describes. The assembled month lands around half a gigabyte on these habits, with one gigabyte as a comfortable planning ceiling for the profile.
The composition explains the plan fit. Marketed pooled plans for video telematics sit in gigabyte units per vehicle. The city van’s assembled half gigabyte sits inside one such unit with room for the exceptional component. A fleet that reads its portal after the first month expects the median vehicle near the assembly and the pool absorbing the spread across the column. The numbers are unremarkable on purpose. The city profile is the baseline the other profiles depart from.
The sensitivity check shows which lever moves the van’s month. Doubling the alarm sensitivity adds tenths of a gigabyte. Doubling the watched minutes adds tenths of a gigabyte. Storing live views by misconfiguration adds gigabytes, the one error that moves the figure by an order, the reason the trade-off page calls it the error the bill reports first.
The long-haul tractor’s month assembles higher on every one of the five components. The floor grows with richer reporting profiles on regulated routes. Supervisory viewing grows because distant vehicles get watched: border waits, night corridors, client checkpoints. The alarm component grows with kilometres and with the wider pre-event windows long routes specify. Pulls grow because the vehicle stays unreachable for weeks, so every investigation crosses the link at recording-tier size. The assembled month lands in the low single-digit gigabytes, commonly one to three on these habits, before any incident month enters the figures.
Coverage holes move the month’s shape without moving its size. Queued parcels from dead stretches arrive in bursts at the next coverage, the delivery pattern the recording page describes. The portal shows quiet days then a spike, the same monthly total a clean route would show spread evenly. A reader who knows the mechanism reads the spike as geography. A reader who does not opens a ticket on a healthy vehicle, the small literacy the monthly ritual builds.
The dual-SIM line splits the same month across two meters. The primary card carries the traffic on its tariff. The backup card idles at heartbeat scale until a hole or a failover, then carries days at full composition. The vehicle’s true month is the sum of two portal lines, the bookkeeping the switching page sets up, with the backup’s near-zero months confirming the failover logic stayed quiet.
Watch discipline is the long-haul profile’s one behavioural lever. Night corridors invite standing watch, and standing watch at the supervisory tier accumulates faster than any other routine habit: an hour a night across a month adds whole tenths of a gigabyte per channel watched. Operations that need standing eyes price the habit into the assembly deliberately. Operations that drifted into it find the drift in the viewing component’s gap, the cheapest finding the reconciliation produces.
Trips that cross month boundaries complicate the reading without changing the totals. A two-week run that starts on the twenty-fifth lands its queued bursts and its pulls in the next month’s column. The per-trip cost is the stable number for costing contracts. The per-month column is the stable number for plan management. A long-haul fleet keeps both views, the trip ledger from the platform and the month column from the portal, reconciled by dates when a number looks odd.
Regulated corridors thicken the floor in a way that belongs on paper before the route is signed. Hazardous and passenger classes carry richer reporting profiles by rule: denser position cycles, supplementary status items, provincial forwarding. The floor component can grow severalfold against the city profile, still inside tenths of a gigabyte, still the smallest mover beside viewing and pulls. The assembly absorbs the regulation as a bigger first line, which is the model working as intended.
The escort vehicle’s month is dominated by one component the others barely use: live windows at full tier. Incident protocols and client contracts put multiple channels on screen at full rate for stretches of minutes to hours at a time. Full-tier streams run at recording-class rates, the consumption page’s upper band, so a single two-hour window crosses gigabytes. A month with several such windows lands in the high single digits to tens of gigabytes, an order above the city van, by contract design and with the contract carrying the cost.
The composition stays the same five parts with one inflated. The floor, alarms and pulls ride along at their usual scale. The live-window component is the plan driver, sized from the contract’s incident assumptions: windows per month, channels per window, minutes per window, at the full-tier rate. The arithmetic is three multiplications on the contract’s own numbers, done before signing, the difference between a priced commitment and a surprise invoice.
The contract clause that follows the arithmetic names who pays for the windows. A client who can open full-tier streams at will is a client holding the fleet’s data budget. The working clause states the included window allowance per month, the rate class beyond it and the report line that counts windows against the allowance. The portal’s daily view supplies the count. The clause turns the article’s largest component from an open risk into a billed service.
The escort profile also shows where 5G enters the series’ map. The multi-channel full-rate window is the workload the 5G timeline page names as a pull, the case that meets 4G’s uplink ceiling first. The monthly assembly stays valid across the radio change. Only the feasible window size moves, the bandwidth question separated cleanly from the budgeting one.
The portal turns the assembly into a checked model on the first invoice. The carrier’s per-SIM view gives each vehicle’s month as one measured figure in one row, the labelling habit from the billing pages making the whole column readable by fleet number at a glance. The first comparison is the median against the assembly: a fleet whose median vehicle measures near its assembled prediction has a model it can plan with for the life of the contract. The gap, where one exists, names its own cause by size, because each component moves in known steps.
The distribution carries more information than the median. A healthy fleet shows a tight cluster around the profile’s assembly, a thin tail of explainable highs, incident months and investigation months, plus near-zero lines on the backup SIMs. The shape is the diagnosis. A wide cluster says profiles differ where configuration should match. A fat tail says the exceptional component is not exceptional. A drifting median says a setting changed fleet-wide, dated by the month the drift began.
The reconciliation closes the loop the assembly opened. Predicted half a gigabyte, measured four: the gap is too large for alarms or viewing, the right size for stored live views, found in the platform configuration within the hour. Predicted two, measured two point three: inside the exceptional allowance, no action. The model puts the measured number against a specific expectation, which is the whole difference between reading a bill and auditing one.
The assembly sheet doubles as the handover document. A new operations hire reads five components, three profile sums and the calibration verdicts, one page that explains every number the portal will ever show them. The sheet outlives staff changes, carrier renewals and platform migrations, because it states the fleet’s behaviour in rates and quantities that survive all three. The same figures settle carrier disputes from the fleet’s side of the table. A bill that disagrees with the portal’s own per-SIM column is a carrier problem with its evidence attached. A bill that disagrees with the assembly alone is a fleet problem with its diagnosis list attached. Keeping the two comparisons separate keeps every dispute short, because each side argues against its own numbers.
The first three months calibrate the model and then leave it alone. Month one finds the configuration errors, the largest gaps with the fastest cures. Month two confirms the cures and tightens the assembly’s quantities to the fleet’s real habits. Month three reads as the steady state: median on assembly, tail explained, backups near zero. The calibrated model then runs for the life of the contract, reopened only by the revisit triggers the parameter pages set for their own numbers.
The top of the monthly column is a short diagnosis list, worked from largest step down. A vehicle measuring gigabytes above its profile with the platform storing its live views has the trade-off page’s configuration error, the largest single multiplier in the system. The cure is one profile setting. The proof is the next month’s figure falling back to the assembly.
A stuck stream is the second pattern: a live session opened and never closed, billing the watch tier around the clock. The portal shows a flat elevated month. The platform’s session log shows the viewer that never logged off. The figure is hundreds of megabytes to gigabytes depending on tier, the arithmetic of a few hundred kilobits left running, caught by any month-over-month comparison.
Roaming days carry their own dates. A border crossing with roaming enabled bills the same bytes on another network’s meter, visible as a dated step in the daily view. The cure belongs to SIM policy, the roaming line the account pages of the carrier contract carry per card. The diagnosis costs one look at the dates against the route log.
A firmware loop is the rare fourth: a terminal repeatedly downloading an update it fails to apply. The pattern is identical large downloads repeating across days, readable in the daily histogram. The cure is a service flag to the maker. The portal catches in days what would otherwise run for a quiet month, one more return on the per-vehicle reading habit.
The backup-SIM column reads as a failover audit on the side. Near-zero months confirm the dual-SIM logic stayed quiet, the expected state on healthy routes. A backup line carrying real traffic names a vehicle that spent days failed over, the cue to read the switching page’s logs for the trigger. A backup line at full month scale names a primary that died entirely, with the vehicle running on its spare since. The column answers in one glance what the switching page’s bench test answers in an hour.
Every outlier diagnosis ends the same way: the next month’s figure returns to the assembly, or the assembly was wrong and gets corrected. Both outcomes improve the model. The fleet’s predicted-versus-measured gap narrows month over month until the column reads as confirmation, the steady state the ritual below maintains.
The plan is sized from the distribution’s upper working edge. A pool sized at the mean leaves half the months over, by construction. The working rule takes a high percentile of the assembled-and-confirmed distribution, the ninety-fifth as a common choice, adds the exceptional allowance, then sizes the per-vehicle pool there. The tail above the percentile is what the pool’s sharing absorbs, the point of pooling.
Pooled plans change the failure mode in the fleet’s favour. One vehicle’s incident month draws from the fleet’s shared headroom, with no per-vehicle cap to breach, so the throttling cliff the billing pages warn about moves from the busiest vehicle to the whole pool’s edge. The pool’s edge is then managed as one number: the fleet’s summed assembly plus the tail allowance, reviewed quarterly against the measured sum.
The plan line closes the procurement question the series opened at the billing page. Per-activity rates priced the behaviours. The assembly composed them into months. The distribution confirmed the months. The percentile sized the pool. Each step is one sheet, one read, one number, the paper trail that lets a finance office approve a data contract without guessing.
A mixed fleet pools across profiles with the same arithmetic. The city vans, the long-haul tractors and the escort vehicles each carry their own assembled month. The pool is sized at the sum of the three profile assemblies plus the shared tail allowance, with the per-profile expectations keeping the monthly column readable whenever the pool’s total number drifts. One pool, three expectations, one quarter-hour read covering all of them.
At renewal the sheet turns into money. A fleet holding two years of confirmed per-vehicle months negotiates from its own distribution: this median, this tail, this seasonal shape, priced against the carrier’s families. The carrier’s first offer is sized for fleets that know none of this. The counter-offer reads off the sheet. The gap between the two is the documentation’s annual return, collected every renewal.
Plan families differ in how the pool behaves at the edge, the one contract read that belongs beside the percentile. Hard-capped pools cut or throttle at the boundary, the cliff the billing pages describe, with queued evidence waiting out the month behind it. Overage-billed pools keep flowing at a per-gigabyte rate, trading a known cliff for an open invoice. Fleets carrying evidence duties price the overage family for the incident months, because a throttled link during an incident costs more than the overage line ever will.
The ritual is one screen and one sort. Open the portal’s monthly per-SIM column, sorted descending by the month’s figure. Read the top ten lines against the four-pattern diagnosis list, largest step first. Read the median against the profile’s assembled prediction. Read the backup SIM lines for the expected near-zero. Note the verdicts in the fleet’s log, three lines a month, dated and signed. The quarter-hour catches the misconfigurations in their first month, the stuck streams by the next read, the drift before it compounds.
Seasonality belongs in the log as a note, with no alarm attached. Holiday routes, harvest contracts and weather months move watched minutes and event rates in patterns that repeat year over year. A fleet’s second year of logs turns those moves into expected shapes, the difference between a number that worries and a number that confirms. The log is the memory the ritual builds, one quarter-hour at a time.
Assemble each profile’s month from the five components before signing anything, on the series’ standing rates and the fleet’s own quantities. Size the pool at the high percentile plus allowance. The mean undersizes by construction. Label every SIM by fleet number in the carrier portal, with backup cards marked as such. Write the fifteen-minute read into the monthly operations calendar, with the diagnosis list beside it. The fleet then knows its number before the carrier does, per vehicle and per profile. Every invoice after the first arrives as confirmation of a figure the fleet already holds on one sheet.
The number comes from assembly, because it follows fleet habits and no datasheet owns it. A city distribution van lands around half a gigabyte on ordinary habits: a small resident floor, occasional supervisory viewing, a handful of alarm parcels, rare pulls covered by the depot card reader. A long-haul tractor runs one to three gigabytes with richer reporting, more viewing and link-crossing investigations, before any incident month. An escort vehicle with contracted full-tier live windows runs high single digits to tens of gigabytes, with the window arithmetic done on the contract’s own numbers before signing. Each figure is five components summed on the series’ standing rates: the resident floor, supervisory viewing, alarm parcels, named pulls and an exceptional allowance.
Because neither controls it. The terminal sets rates per activity. The fleet’s habits set the quantities: watched minutes, event counts, investigation pulls, contracted live windows. The same terminal in two fleets produces two different months. Equipment pages quote storage and channels, carrier pages quote pooled gigabyte units. The per-vehicle month sits between them as the fleet’s own arithmetic, assembled once from the five components, confirmed in the first quarter and checked against the portal monthly after. The omission is structural, the reason the assembly method exists and the reason it transfers across carriers and platforms unchanged. The assembly sheet doubles as the handover document for staff and renewals.
Four patterns cover the column’s top. Stored live views from a platform misconfiguration add gigabytes, the largest single step in the system. A stuck supervisory session bills its tier around the clock, hundreds of megabytes to gigabytes by the month’s end, caught by any month-over-month comparison of the same vehicle. Roaming days bill the same bytes on a foreign meter, visible as dated steps in the daily view against the route log. A firmware download loop repeats identical large transfers across days in the daily histogram. Each is readable in the portal and each has a one-line cure. The proof of every cure is the next month’s figure returning to the assembled expectation.
From the distribution’s upper working edge. Take a high percentile of the confirmed per-vehicle months, the ninety-fifth as a common choice, add the exceptional allowance, then size the pooled plan there. The mean undersizes by construction, leaving half the months over the line it drew. Pooling absorbs single-vehicle incident months in shared headroom, moving the throttling risk from the busiest vehicle to the managed pool edge. Evidence-carrying fleets price overage-billed pools over hard caps, because a throttled link during an incident costs more than any overage line. A mixed fleet pools across profiles, sized at the sum of the profile assemblies plus the shared tail.
No. Queued alarm parcels from offline stretches upload after reconnection at the same sizes and with the same event times, on the resume mechanism the recording page walks. The month’s total stays as the route’s behaviour set it. The delivery pattern changes: quiet days then a burst at the next coverage, the spike a portal reader learns to attribute to geography before opening a ticket on a healthy vehicle. The monthly figure stays inside the assembly either way, with only the daily shape moved by the route.
One screen, fifteen minutes. The per-SIM monthly column sorted descending, top ten read against the four-pattern diagnosis list, median read against the assembled prediction, backup SIMs confirmed near zero as a failover audit, verdicts logged in three dated lines for the seasonal memory. The habit catches misconfigurations in their first month and turns every later invoice into a confirmation of a number the fleet already knew. The first three months calibrate the model. The years after run it, with two years of confirmed months becoming the fleet’s own negotiating position at renewal.
