bbeesley
VIP
join:2003-08-07
Las Vegas, NV
| reply to dvd536
Re: [AZ] Upload results
Speedtest servers should always be taken with a grain of salt because there is so much variability that may artificially affect the measurement.
Things like:
•varying network paths between you and the test that may have varying degrees of congestion...your route could even traverse multiple foreign networks
•interface capacity of the test server - does the server have something like a 100mb Ethernet port, or a gig port, or something else altogether?
•CPU, memory, drive speed of the server? does the server has insufficient resources to run the multiple requests?
•other applications on the server...is the server running multiple duties and if so, what are the service priorities?
•number of other users making the same request at the same time as you are...if the server is popular at particular times, this could skew the results
•is the server or it's network currently being attacked by something like a DDOS...for most networks this is a now common event that may not take the server down but it will affect performance sometimes to a measurable degree
•your PC or the code used for the server...if your PC is busy or if the code for one server is more efficient than another, you will get variable results
•anything else I might have forgotten
as you can see there is really no way to accurately use a server that is off net to do anything than acedotely measure your service. In order to have an accurate test, we would to place equipment at your location and it the CMTS to measure you and only your connection, eliminating your PC, the server, and the plethora of networks in between.
or a shorter answer would be - "your mileage may vary" |
