How to perform an accurate speed test.
When performing a speed test, you should know there are many factors which can distort the results. To ensure you receive the most accurate results, you should apply these cardinal rules by following the below steps.
- Choose a speed test Service which allows ample bandwidth to their speed test servers. For example http://speedtest.mybroadband.co.za/.
- On the test itself, make sure you test with the closest server to your physical location. For example, if you live in Emalahleni it would be better to select a Johannesburg server than a Durban one. The server with the shortest latency & appropriate bandwidth will yield the most accurate results.
- Make sure your computer is connected via an Ethernet cable and NOT with Wi-Fi. This is to help ensure no other devices or people on your local network are using bandwidth while you perform the test.
- Triple check you have no other programmes or unnecessary processes running in the background, such as windows updates, iTunes or antivirus. These will draw bandwidth from your line while running the speed test and obscure your results.
For best results…
If you are a BitCo customer and you would like to perform a speed test, we recommend http://speedtest.mybroadband.co.za/. This is because MyBroadband is a NAP Africa peering partner, as is BitCo. Our servers are literally stored in the same data centre, meaning you will obtain a bona fide reflection of your speed (provided you have adhered to the rules above, specifically 3 & 4).
We have also verified that the MyBroadband speed test servers allow generous bandwidth, so that multiple tests can be run at once without anyone’s results being skewed.
Remember that a speed test to a server which is only has a 100Mbps connection to the Internet (ISP) will most likely not give the correct reading when you are testing your 1Gbps connection. This is also true if there is a large quantity of test being performed for smaller connections.
For example, you are only testing your 10Mbps connection, there may be another fourteen users testing their 10Mbps connections at the same time. All of these will pull from the 100Mbps server connection and thus affect each user’s result. This is because the capacity of the speed test server is not optimised for 15 users all running tests at the same time.
International speed tests
International servers are also limited by the amount of bandwidth they are able to receive and send. It is therefore unlikely that you will achieve the same results on an International speed test as you would from local servers.
Below are some examples of how easily your results can be distorted by not following the guidelines above.
The control used was 100Mbps synchronous (100Mbps download and 100Mbps upload) BitCo Fibre connection at the BitCo Headquarters in Johannesburg.
Ethernet cable versus Wi-Fi
Results from connection during normal office use, between 8am – 5pm during the week with multiple users on the network, drawing bandwidth.
The True Test
Best results – Here, the live IP of the office was placed onto the laptop and the office firewall removed. All bandwidth available was dedicated to the laptop, showing the full speed. Remember to keep in mind that a laptop network card is only 10/100 and 97.73Mbps would be considered full speed.