With the advance in internet connections and their speeds, a lot of people are now wanting faster and faster connections, with fibre connections offering speeds from 10MB to 1,000MB. In addition to this, people also want value for their money, and so want to make sure that they are getting what they are paying for.
A common way that a lot of people try to check is by using online speed tests, such as speedtest.net. However speed test sites like this do not always result in fair tests of the connection. This is because from where you sit at your computer, to even where you connect to your high speed connection, there is a multitude of hops, or bottle necks, including everything from your own PC, to routers, switches, hubs and firewalls before you even connect to your high speed connection.
From here, the high speed connection is between this last point and what is called your Point-of-Presence to the internet. The high speed connection runs here. To get a fairer test of your speed you need to be in some way connected to whatever the last step in your network is, whether it’s a firewall or router or something else, in some cases things like Draytek routers or Sonicwall routers, and then test the speed from there, removing the rest of the hops, stopping them from slowing the connection.
Think of your Point-of-Presence as the point at which the slip road connects to the motorway. The motorway is the internet and the slip road is your connection to the internet.
As well as that, there is also the issue of traffic on your network. If you have multiple users using your high speed connection, then this will bring down your own speed. If you have a 50 MB connection, and ten users, each using 5 MB, then this will throttle your connection speed down.
However, this is only but one part of the problem with these speed tests. The other part of the problem comes from where you are connection to on the other end as this is the other end of the test, where the different speeds are tested.
The first issue is the location of the server/machine. If the machine is located physically further away from you, then logically, there are more hops in the connection. For example, if you are running the test in Edinburgh a server located in Poland will result in a slower speed, as for one there are a lot more hops to get to the Polish server than there would be to hit a server in another Edinburgh based company or ISP. The Edinburgh server may only be about five hops away, whereas the Polish server could be fifty-five!
Secondly, there is also the issue of the connection on the server end. If this connection is slow, then the responses back from the server will be slower and give a lower speed for your test. If the server is located in a remote area such as the Outer Hebrides then it is likely to have a much slower connection than one in a city such as Edinburgh or Glasgow.
Finally, it is incredibly unlikely that you will be able to connect to a server in your Point-of-Presence. The way to get the best results on oyur speed test is to connect from your last hop in the network, a firewall/Sonicwall device, to the same point on the Point-of-Presence’s network. This would be the fairest test of the high speed connection and give ou the best indication possible of what the actual speed of your connection really is. But the likelihood of getting a server at your Point-Of-Presence is very very low. Often the best you can hope for is a server in the city, or even in your country in some cases.
To put it simply, is it possible to drive from Edinburgh to London at seventy miles per hour the whole way?
No, because there is other traffic and different sizes of roads between here and there. The internet works the same way.