“What’s the best way to measure my DSL speed? Am I getting the speed I paid for? These questions are being asked to me a lot of times so I decided to write something about it. Note that i’ve made a few references to PLDT because it is my current provider, but technically these tests will apply to any ISP.
There are a couple of Internet speed measurement websites around. I primarily use 2, SpeedTest and TestMyNet. I use SpeedTest most of the time since it supports local speed tests, which I will explain shortly. I only use TestMyNet to troubleshoot some issues I have with my connection. To measure true Internet speed, I use a mix of regular download and torrent download.
There are 2 speeds you have to be aware of with regards to your residential Itnernet connection. First is the speed from your home, to your ISP’s data center. I call this “local speed” and my own personal criteria is that these should be close by a margin of -5%(low) to +15%(high) from your DSL plan speed. Example: You are on PLDT Plan 1299(2Mbps, 2013), you should be getting 1.7Mbps-2.1Mbps local speed. Speedtest.net have local servers for most of the major PH ISPs. A local server is a speedtest server hosted by your ISP. Speedtest allows ISPs to host their own server so their subscribers can measure the speed from their home to their data center. So if you open speedtest.net, pick the server of your ISP. Below is my local speed at Plan 3000 using a server hosted by Smart(PLDT), this goes from 8Mbps to 9.9Mbps depending on the time of day.
Now, what’s affects local speed? Primarily, DSL signal degrades over distance. The farther you are from the ISP, the lower your DSL signal is and proportionally, the higher your DSL signal noise. Also, if you have worn out and rusted cables, grounded lines and the likes, this will affect your local speed. From inside your home, this is something that you can check and verify. Check the PLDT drop lines outside your home, is it properly grounded? Check the junction box – are the contacts free from rust? Check the splitters etc. So before calling your ISP to bitch and moan that you aren’t getting your speed, make sure you have those items checked.
The second type of speed you need to be aware of is your actual “Internet speed”. This is the speed you get when you download files off the Internet. Compared with local speed, the files you’re getting to test Internet speed is beyond your ISP’s data center.
There are 2 ways to check your Internet speed, see below. I primarily use Linux disk images(from centos.org) as downloads since they’re hosted on large data centers
1. Use a download manager(I use the one from freedownloadmanager.org) and download a file from a good source. A download manager creates multiple connections to a site to download the file. Here’s my average download using a download manager at my speed:
This translates to roughly 9.1Mbps, which is pretty close to my “local speed” from speedtest.
2. Second way to test is to download a file using bit torrent. Bit torrent is a file sharing protocol(when you download a chunk of a file, you also upload a chunk of it to others), thus, the more people downloading a file, the faster your download speed will be. It’s like using test #1 but this time you’ll have multiple download sources(see below, i’m downloading from 56 sources!), thus your speed is highly maximized. Here’s my torrent speed – roughly 9.9Mbps.
So to sum up:
- Use speedtest.net to get an overall view of your speed(local speed). When you encounter sluggish internet, go to speedtest.net first and check if your speed is down from the usual readings. If it does, it might be a physical issue or a local downtime perhaps. If it doesn’t, it might be a DNS issue or the site is having problems. Register for an account on speedtest.net so you can have a history of all your speedtests.
- To measure your real internet speed, download a file using a download manager or using bit torrent. This is what matters most since this is the speed you are paying for.
- In any troubleshooting task, isolation is the key. Make sure you have reliable local speed when testing Internet stability. When measuring actual throughput, rely on your Internet speed more than local speed.