by Geoff Courts - @Macnamara_Geoff
If you are experiencing a poor internet connection, there are a number of websites that offer to test your upload and download speed for you. The problem is, if you want to get a truly accurate result, the act of measuring can actually affect the result. In physics this is called the Uncertainty Principle; this states that, on a quantum level, you cannot know both an objects’ speed and location, because the act of measuring will impact on the outcome (all measurement exerts some force or energy of the object being measured).
On a macro scale, the energy impact of photons of light has a negligible effect, but when measuring the velocity and location of individual particles, this can give experiment tainting results.
Speed test websites that use heavy flash simulations to display moving dials and pictures are consuming bandwidth at the very time you are trying to get an accurate reading of the connection you are using. Similarly, it’s no good testing the connection if someone else in the office is streaming BBC iPlayer, and someone else is downloading a large PDF report off a website. Uploads and downloads will affect directly the results of the speed test, including downloads from flash animations.
http://speedtester.bt.com/ gives an accurate reading of the actual download speed of the connection. It requires only that the PC you are using is connected directly to the router on the connection, and that no other machines are connected on the network, so the line is as clear as it can be and the externalities of other connected devices are avoided entirely. (It does require a Java plugin, however)
What’s more, the test results are logged with BT and, in Macnamara’s case, Zen (our preferred ISP), allowing them direct access to the results in case a fault is raised on the line.
If the test fails, or returns ‘inconclusive’, a second test may be initiated, which involves re-setting the router username and password to speedtest@speedtest_domain. This connects directly with the BT exchange, giving the truest reading on the line possible. If the result is poor, then the evidence is strong that the issue lies with the line itself. If the performance is shown to be good, it suggests that there may still be another network issue at play, and further investigation may be required.
No comments:
Post a Comment