@AE_Collector wrote:Sounds reasonable but that doesn’t answer whether there is a set amount of data moved to see how long it took which would not result in higher data consumption if 4G rather than 3G OR if a set amount of time is used to move as much data as possible within that time frame which certainly would consume more data than the first method.
The data itself is a bunch of random garbage files with standardized block sizes, fetched from the speedtest servers then sent back, arranged contiguously because read/write and transmit/receive operations are always fastest for sequential data access.
The speedtest app doesn't care how much data it uses - it will hog all available memory and all available bandwidth, it just deletes the temporary mess when it's done. The "set amount of data moved" is "as much data as the device and the network can move during the test".
If you are connected to a fast network and your device has fast processing then it can consume a lot of data in a single run. It's trying to move impossibly huge data before cutoff, but the bottleneck from the network or the device will impose a real limit and thus a real performance result.
When I run a speedtest from a home computer with 10Gbps optik it downloads almost 70GB and uploads almost 24GB because the network bandwidth can support it and the computer can process it. And this represents less than 10% of the actual packet requests sent by speedtest, it would consume 1024GB if the hardware could sustain a fast enough rate before cutoff.
Mobile devices are not as well-equipped for large data loads. Wireless networks are pitifully slow and low bandwidth rates compared to wired networks. But speedtest can still gobble up most of a GB per run on a decently powerful mobile device with uncapped LTE data access, good signal, and low network traffic conditions.
