2

I'm wondering how it's possible that download speeds vary so much depending on location.

Here are test dedicated servers:

  1. https://10gbps.io/datacenters they offer various test download files. They are connected to 10 Gbps Internet. (I only tried European servers.)
  2. One dedicated server located in Paris, France http://www.iliad-datacenter.com/ connected to 1 Gbps Internet offered by scaleway.com.
  3. One dedicated server located in Paris, France connected to 1 Gbps Internet offered by online.net.

Here are test computers:

  1. Windows 10, located in Warsaw, Poland (upc.pl is the ISP).
    • An online speed test shows 120 mbit/s.
    • The download speed from all three test dedicated servers reaches ~14-15 mb/s at all times (that's the max speed for this Internet connection).
  2. Windows 10 located in St. Petersburg.
    • An online speed test shows over 100 mbit/s.
    • The download speed from all three test dedicated servers reaches max ~2 mb/s (sometimes jumps to 5mb/s for a short while) at all times.
    • This one is optic fiber and gives 1ms ping and very stable connection. Also, downloads from Nvidia/Apple/Microsoft websites (and other similar) are reaching max speed of 10+ mb/s at all times.
  3. Windows 10 located right next to Warsaw, Poland but with different (small, local) ISP.
    • An online speed test shows over 100 mbit/s.
    • The download speed from all three test dedicated servers reaches max 2mb/s. Similarly from Nvidia/Apple/Microsoft it's always full 10+ mb/s. It's optic fiber, ping is always 8-9ms.

So, how come I get very limited download speed from two locations and at the same time I get full speed from one?

1 Answer 1

0

Sometimes there may be a bottleneck somewhere along the network path between two systems. If you use the traceroute command - tracert on Microsoft Windows systems, you can see the network hops between a source and destination system. E.g., if you issue the command tracert example.com you may see that there are 10 hops between the system where the command is issued and the destination server example.com. Those intervening systems may be routers, firewalls, or other pieces of network gear. You might then try the same command for anotherserver.com, i.e., tracert anotherserver.com and see 15 hops in between, because a different network path is used. The longer path might actually be a quicker path, though, if one of the pieces of network equipment in the path to example.com is overloaded with other network traffic.

Also, to understand where the limit may be on the bandwidth from point A to point F, you may need to know the speed of the connections between intervening pieces of network equipment. E.g., if the path is from A to B then C, D, E, and finally F, if A is connected to B at 100 Mbs and E is connected to F at 100 Mbs, but C has only a 10 Mbs connection to D, then the maximum bandwidth you can achieve between A and F will always be less than 10 Mbs unless there is some other network path available between the two endpoints.

It is also possible that the server at example.com is experiencing significantly more network traffic than the anotherserver.com system at the time you test. Or that system may not be as powerful, e.g., it may have a slower CPU or other hardware constraint that makes it less able to cope with the demands placed upon it.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .