By Charles Miller
I was called to troubleshoot a woman’s broadband internet connection because she was complaining that her connection speed was slow. When I arrived she showed me a printout one of the popular internet speed test sites found online and told me it reported her service as being a “Grade F” for failing. I asked for a few minutes to run a different speed test that took longer.
While I was working she continued using her own laptop and I could tell she was running a quick speed test over and over and over and over and over. Suddenly she jumped up and called out “Look! Look! I told you, it’s slow!” I could see the reportedly-slow test results, and asked her “What were the results of the previous test, and the ones before that?” Silence. She did not want to admit that several previous tests showed her internet speed was normal, she only wanted to focus on the one test that reported the slowest.
I explained that online traffic is inherently bursty. When you load your email page you receive a burst of data that stops when the page is finished loading. When you click on an email you receive another short burst. The entire internet works in this way and should you happen to run a short speed test while your connection is busy handling another burst of data then your test results will show your speed was indeed very slow at that instant. She continued to insist “It’s slow!”
Her test proved there was a point when the connection was slow for a few seconds, but that does not necessarily mean it stayed slow. In all likelihood, the speed was slow for a few seconds, speeded back up for a few minutes, slowed down again for a few seconds before going back up again, and so on. That is simply the characteristic of how measurable internet speeds fluctuate.
I then showed her the test results I had from the different test I had run that measured over a period of several minutes and which showed her average speed sustained for that time period to be quite fast, in fact everything she was supposed to be getting. She was still unconvinced and ran her short burst test over and over and over and over and over until she finally got another result to show me. Disregarding all the tests that showed her connection to be okay, she showed me the one result that proved her claim: “It’s slow!”
Then she challenged me on why my test results were accurate and hers were not. The answer is that the popular speed test sites compute and report your bandwidth as an estimate based on downloading and uploading a very small file from their server to your computer. A more accurate measure of actual bandwidth requires uploading and downloading a much larger volume of data continuously for a longer period of time. The longer the test, the more accurate the results. Several minutes is okay, repeating that longer test several times and averaging the results would be even more better; but who has the patience to wait an hour for the results?
The results of speed tests are rarely the same, even the exact same test repeated. There are so many factors that affect test results.
Charles Miller is a freelance computer consultant, a frequent visitor to San Miguel since 1981 and now practically a full-time resident. He may be contacted at 415-101-8528 or email FAQ8@SMAguru.com.
