SpeedOf.Me draws the whole measurement as it happens: every sample file, every ramp, every stall. That makes it a teaching instrument as well as a speed test. Below is a lab exercise you can assign as it stands, an answer key for the questions it raises, and standing permission to put our screens in your own materials.
Open speedof.me and start a test. For the first run, tell students to ignore the final figure entirely and sketch the shape of the graph instead: where it rises, where it flattens, where it drops.
Each band in the graph is one sample file, and within a band the speed starts low and climbs. Ask the class why a larger file measures faster than a smaller one on the very same connection.
Repeat at the far end of the building, then next to the access point, then wired if the room allows it. Same line, different results. Ask which layer of the stack changed, and how they would prove it.
Run a multi-stream speed test alongside ours and compare the totals. Ask which of the two numbers a single video stream is going to get, and why a provider might prefer to advertise the other one.
A shorter version for one class hour: steps 1 and 2 only, then the first three questions below. The full four-step version runs to about ninety minutes with discussion.
TCP does not begin at full speed. It starts with a small congestion window and grows it as acknowledgements come back, roughly doubling each round trip. A 128 KB file is finished before that window has opened, so it reports a fraction of the real line rate. The first few bands of the graph are slow start, drawn to scale. On a fast connection with a long round trip, students will see several bands go by before the curve gets anywhere near the advertised speed, which is the single most useful thing in the exercise.
The window keeps growing until something other than the protocol becomes the limit: the access line, the Wi-Fi link, a congested hop. The flat top is that limit. If a run never flattens, the test has not yet reached a file big enough to find it, which is exactly why the sample sizes keep doubling.
SpeedOf.Me keeps stepping up until one sample takes longer than eight seconds, and reports that sample. Eight seconds is long enough that the slow-start portion is a small share of the transfer, so the number reflects steady state rather than the ramp. Ask students what would happen to the reported figure if the test stopped at one second instead.
Several connections share a bottleneck more aggressively than one connection does, so their sum can approach the line rate while any one of them does not. Both numbers are real and they answer different questions. A speed test that adds streams together tells you what the line can carry; a single stream tells you what one download, one video or one upload is going to get.
Speed is quoted in megabits per second and file sizes in megabytes, and there are eight bits in a byte. A 100 Mbps connection moves about 12.5 megabytes a second. Students who miss this will conclude the test is wrong by a factor of eight, which is a useful mistake to make once, in a lab, rather than in an argument with a provider.
Latency is the round trip to the server and back, measured over ten samples with the lowest reported. Jitter is how much those ten varied. Ask which of the four numbers on the screen a video call cares about most, and why a fast connection can still be a bad one.
Method, sample sizes and server selection in full: how the test works, and why we call the result Real-World Speed.
Standing permission for teaching. You may reproduce screenshots of SpeedOf.Me in course materials: slides, handouts, lab manuals, worksheets, LMS pages and exam papers, including materials your institution charges for. Two conditions, and no paperwork: do not alter the figures or labels the screen shows, and credit it as "Courtesy of SpeedOf.Me" with the address speedof.me. This covers the screenshot only. It is not a licence to use our name or logo to promote anything.
Publishing in a textbook, a study guide or another commercial work? Write to contact@speedof.me first. We grant that too, and asking gives us the chance to check your annotations against what the test actually does before the figure goes to print.
Each of these is that publisher's own editorial choice, not a partnership. The full list, with a link to every source: who cites SpeedOf.Me.
A different job from the lab above: checking that a room full of devices can actually do what the lesson needs. Programs that run assessments in schools already use SpeedOf.Me this way, telling staff to test classroom connectivity while students are on their devices rather than after hours, because an empty room measures nothing useful. University help desks point students at it for the same reason, as part of a pre-class or pre-exam connection check.
If you would rather run that check inside your own portal than send people to another site, the same engine is available as an API: one script tag, and the download, upload, latency and jitter figures come back to your code. Bandwidth for testing is included, and there is a free tier.
Using this in a course and want something we have not put here, such as a printable worksheet or a figure at a particular size? Write to contact@speedof.me.