← Back to the chapter · Next detail →
Technical detail · NB-IoT Technical Specifications
Transaction timing and falsifiable claims
Suppose a 200-byte application payload completes in 4.0 s in good coverage and 18.0 s with repetitions at the edge of service. Application throughput is for the first transfer and about for the second. Neither is a universal NB-IoT rate: signalling, retries, scheduling and radio conditions shape the transaction.
Send the same 200-byte fixture repeatedly and report a distribution rather than only the fastest result. Peak values remain useful radio references, but planning needs measured transaction duration and server receipt. A successful radio attachment does not establish that the application received the reading.
Pair each specification claim with a falsifier—an observation that would show the claim does not hold. Examples are a payload exceeding its active-time budget, a weak-site trace with repeated retries, a timer grant breaking command reachability, or an operator profile lacking the required feature.
Predict the loop before testing. An unsupported band should make registration fail visibly. Added attenuation should change repetitions, delay or energy. A requested power timer should be compared with its grant in the modem trace. Include attachment, retries, response windows, recovery and return to sleep in the power measurement.
Use the worked transfer as a method, then verify each promise on the deployed combination. Keep the acceptance limit and observed exception together so a reviewer can see both what passed and what must be retested.