|
Even if you had fibre to your house in most cases (if your within a couple hundred metres of the node) it is not going to make any difference to the speed you receive. The issues, as always, is network design. Every network is designed to handle a certain throughput, not to max out every possible connection speed. The devil is in the muxing design, how many user circuits are competing for high speed packet space. The exchanges are designed for a certain 'arrival rate', and the equipment has a service capacity that is based on a model. Standard queuing theory stuff but at a high speed. Some people may get bufferred temporarily until packet space is available, sometimes there are timeouts and discards requiring retransmit. All if this affects the speed you see. Speedtest does not map the intermeadiate routing, you get no information about how many hops your test has made. Each hop is a potential delay factor. The use of mostly capital city test servers is likely to disadvantage regional or country testers, but in the real world being outside a major city can have session advantages due to city traffic congestion. When my daughter wants to fight for popular concert tickets on Ticketek, she gets me to login and establish a session from my place. I have no problems getting a session, even on the day Adele tickets went on sale. Her problem is she is queuing for space in city exchange hops whereas i have almost direct routing (two hops as far as i can tell) and there are less competing for space. So there are many factors affecting your performance.
|