The latest MyBroadband Speed Test statistics show that Axxess, Cybersmart, and MWEB offer the lowest local latency on ADSL connections.
When looking at 20Mbps and 40Mbps VDSL connections, MWEB, Cybersmart, and Internet Solutions reigned supreme.
MyBroadband’s speed test service uses Ookla’s platform, and is hosted in Teraco’s vendor-neutral hosting facilities in Johannesburg and Cape Town.
Through NapAfrica, all network operators present at its peering points are provided with a free 1Gbps connection to the MyBroadband speed test platform.
To ensure that only valid results are recorded, MyBroadband filters speed tests based on network information from the service providers.
The following tables show how different Internet Service Providers (ISPs) in South Africa performed on ADSL, VDSL, and fibre.
| ADSL ISPs | ||||
| ISP | Connection | Download speed (kbps) | Upload speed (kbps) | Latency (ms) |
|---|---|---|---|---|
| Axxess | ADSL | 3,677 | 537 | 52 |
| Cybersmart | ADSL | 2,973 | 433 | 54 |
| MWEB | ADSL | 2,762 | 439 | 58 |
| Telkom | ADSL | 3,002 | 490 | 62 |
| Internet Solutions | ADSL | 4,092 | 532 | 65 |
| Afrihost | ADSL | 3,145 | 469 | 68 |
| MTN Business | ADSL | 2,157 | 1,640 | 191 |
| – | ||||
| VDSL ISPs | ||||
| ISP | Connection | Download speed (kbps) | Upload speed (kbps) | Latency (ms) |
| MWEB | VDSL | 15,876 | 1,170 | 30 |
| Cybersmart | VDSL | 21,687 | 1,783 | 34 |
| Internet Solutions | VDSL | 18,609 | 1,752 | 36 |
| Afrihost | VDSL | 11,484 | 1,290 | 39 |
| Axxess | VDSL | 9,484 | 1,373 | 41 |
| Telkom | VDSL | 19,048 | 1,781 | 44 |
More on broadband
Gauteng to have 100% internet connectivity by 2019
Fibre, ADSL and Mobile: a rating of local broadband services
Loading ...
Another article with a very poor method (really, using speedtest results to MyBroadband speedtest servers to represent gaming latency to servers that are mostly not hosted at Teraco DCs) and totally useless statistical information.
If you are going to use different sample sizes it is critical to indicate the sample size. Even if the sample sizes are the same, you should be providing a confidence interval, and some justification for which statistical distribution you used.
Please get someone on your staff to attend Statistics 101 or at least ‘Statistics for dummies’ before peoviding averages of unknown sample size as fact.