A bandwidth test is a program that sends one or more files of known size over a network to a distant computer (for example, your own computer), measures the time required for the file(s) to successfully download at the destination, and thereby obtains a theoretical figure for the data speed between two or more points, usually in kilobits per second (Kbps) or megabits per second (Mbps).
Bandwidth test results vary greatly, even from moment to moment, and occasionally produce absurd or improbable figures. Factors that affect test results include:
- Internet traffic (speed generally decreases as volume increases)
- Variable propagation delays (can artificially inflate or degrade the result)
- Noise on data lines (has a real detrimental effect)
- The size(s) of file(s) used for the test
- The number of files used for the test
- The demand load on the test server at time of test
- Geomagnetic and/or thunderstorm activity
In order to get a reasonable estimate of bandwidth (sometimes referred to as throughput), experts suggest that three or more different test sites be used, and that each test be conducted six times at each site. Then the top and bottom 1/3 of the figures should be disregarded. Finally, the middle 1/3 of the results should be averaged.