Virgin Media questions accuracy of broadband speed-testing

by Scott Bicheno on 14 October 2008, 15:09

Tags: Virgin (NASDAQ:VMED)

Quick Link: HEXUS.net/qapq7

Add to My Vault: x

Yet more issues

Virgin Media's position

  • We know customers want to check their speeds to ensure their line is running as expected, but the information they rely on must be accurate - currently it is not, yet many web-based speedtest sites make judgements on broadband performance using what is effectively dirty data.
  • Most web-based speedtest sites are also broadband comparison sites or re-sellers. Speedtest results often average out each ISPs result, meaning there is often no distinction between faster packages and slower ones - this is too simplistic.
  • We are keen to work with all speedtesters - hardware and software - to help improve industry standards in reporting of broadband performance.
  • Just as we share the consumers' view to have clear advertising of speed and performance in broadband marketing, we feel the industry needs to help consumers understand any limitations of different testing methodologies and agree some common standards to ensure the information customers are relying on from comparison/switching sites is as accurate as it can be.
  • We have launched an initiative whereby we are inviting all main speedtest sites to work with us to help improve their services ahead of the launch of our 50Mb service.

 

Do you think Virgin Media has a point or is this just an attempt to pressurise the speed-testers into doing things their way? Do you use speed-testing websites and if so how reliable do you find them? Let us know in the HEXUS.community.

 



HEXUS Forums :: 14 Comments

Login with Forum Account

Don't have an account? Register today!
While I can see where Virgin are coming from in terms of test packet size and test server load/bandwidth I find the blaming of general internet congestion to be a bit odd after all if it effects 50Mbit services it does the same to 2Mbit so everyone is going to be effected. And after all do you care that your line can theoretically transfer at 50MBit or that when you download from a real server somewhere that it is fast? While I can also see that people want to get what they pay for, the current speed tests probably provide a good idea of what you can expect to achieve in terms of everyday internet use.

If server bandwidth congestion and packet size are an issue perhaps you should conduct multiple (10?) speed tests with different sites simultaneously and compile the results?
I wonder how many speed tests you can run at peak times (most of the day) before you get your bandwidth capped?
So how many minutes can you download at 50Mbps before triggering the download throttle?
To test your VM use thier internal FTP…. ITs a shame the rest of the internet is not on their FTP! If they can't give everyone even decent speeds at 10Mb why offer 50Mb?

Also I think its about time VM looked at how their speed is delivered, pings are all over the place, loss is a an issue at peak times etc etc.
Steve
So how many minutes can you download at 50Mbps before triggering the download throttle?

The speed has more than doubled but I'd say they'll increase the throttle limits linearly, so you'll spend more time not on your 50mbit internet :)

They say they should be testing speeds to their servers, but what about the real internet where not everything is on their servers. It's not the speed to their servers they should be worrying about it's the average speed they can provide to most servers and the pings. The faster speeds you can provide on a majority of servers makes the ISP better even if it can't reach full speed.