10 November 2020

All About Speedtests [Different Servers, Different Results?]

As quarantine continues and online classes and work from home is the new normal, a reliable internet connection is critical. One of the ways of checking if your connection is good is through speedtests.

In Youtube, you cannot see a video about WiFis without these speedtests. Everyone wants to check if Globe or Smart is the fastest, everyone wants to know which router is the fastest. You'll never run out of speedtests comparison videos in Youtube! 

On speedtests videos you'll always see some comments alleging dishonest speedtests only because of the server used. Most are assuming that if you are testing the speed of a PLDT modem then your server should be PLDT or Smart. If your speedtest server is Globe then maybe they are assuming that Globe can manipulate the test and somehow give slower speeds.

Let's put that to the test: So with my Greenpacket D2K modem, what I did is to check different servers to see the differences in the speed.

Based on Ookla's website: At the start of a test, Speedtest automatically chooses a nearby server with a fast ping result to measure the maximum potential of your internet connection. 

The automatically chosen server in my location is Cablelink Paranaque. I then compared this to different servers in 3 locations, Manila, Makati and Cebu using servers from PLDT, Smart, and Globe on each location.

In Manila, PLDT showed the fastest download speeds and Globe with the slowest. This is the opposite in Makati and Cebu where Globe showed the fastest speeds and PLDT showed the slowest. But one thing I noticed in all tests is that Globe always had the highest ping. Ping is the reaction time of your connection–how quickly your device gets a response after you’ve sent out a request.

Now that we've shown that the speedtests done on competing networks does not necessarily mean that it will be slow, the differences in all servers are noticable. This is because Ookla's speedtest measures your real-time network connection, variations on tests taken minutes of each other will be based on netwrok congestion and available bandwith. And with a LTE based connection, this should be expected.

I tried doing speedtests four consecuctive times with the same server and the difference between the slowest and fastest download speed is 7.32Mbps which just shows again that speeds does not necessarily depend on the server used, you'll also get different results even on the same server because the the speedtest measures your internet speed for that specific time.

Just as a reference, Ookla recommends at least 2mbps for emails and social media, at least 10mbps for video chat, and at least 25mbps for HD streaming. But in the end I believe that it is more important to look at your experience rather than judge your internet based on speedtests. While it would be nice to have fast download speeds, it wont matter if you get disconnected regularly anyway. There are also a lot of factors to consider such as how many are connected to your wifi when you do the speedtests and speeds would even also depend on the band use hence somtimes the need for admin access for band locking.

With that I hope this helped you in understanding a bit about how speedtests works!

No comments: