大象传媒

bbc.co.uk Navigation

Darren Waters

Speed Diary: Day One

  • Darren Waters
  • 17 Feb 09, 15:10 GMT

I am one of the fortunate; a member of the broadband elite.

Last week I had a Virgin 50Mbps broadband connection installed and the goal is to put it to the test over the next week or so and explore the limits of the online fast lane.

Yesterday I described my set-up experiences but I wanted to focus Day One proper on raw speeds, and putting the connection through an obstacle course of downloads and uploads.

The first place to start is with some formal speed tests. Virgin's own website says that just one third-party site is equipped to deal with its XXL connection, .

I also had a link to an internal Virgin speed test site.

Screengrab of wire speed test

I tested the connection on both a wired and wireless configuration. (For those who want to know: an 802.11n router with WPA2 encryption)

Wired

Thinkbroadband put my download speeds at 37Mbps downstream and 1.5Mbps upstream.
The Virgin test put my speed at 46Mbps down and 1.5Mbps.

So why the variation? Andrew Ferguson, from Thinkbroadband.com, told me it was to be expected. The Virgin test had "less network" in the way between my connection and its own test.

Wireless

Both Thinkbroadband.com and Virgin put my download at about 35Mbps and upload at 1Mbps.

If you are seeking absolute speed then an Ethernet connection, always beats wi-fi - even the 802.11n standard. But even so, my contact at Virgin felt I should be seeing faster speeds over the wireless connection. I'll be keeping an eye on this speed over the week.

But there are plenty of other reasons why you won't always see the advertised speeds - even if you are on a wired connection.

The age of your computer, interference from other devices, bandwidth saturation on wi-fi can all have an effect.

So what about some real world tests? How would the connection perform under a heavy load?

The first thing I tried was a straight single download of a large file - in this case, the Linux distribution (an open source operating system) Ubuntu. This is a large file of almost 700MB and I got speeds of up to 5.4MB per second during the download and it landed on my desktop in about four minutes.

Screengrab of Ubuntu

One of the best ways to test your download is to simultaneously download a series of large files and then add up all the downstream speeds. I downloaded five large game demo files and each one was flying down the pipe at more than 1MB per second.

Screengrab of downloaded game demo files

These two tests would seem to indicate that I was hitting very close to the advertised 50Mbps.

It's not that often I download a large program file but I do download and stream quite a bit of video. So how would the connection cope with HD video?

I picked a HD version of 30 Rock to download - in all, about 750MB of comedy heaven. I went from click to buy to click to play and Tina Fey in HD quality in under six minutes.

HD trailers streamed in 720p from the Apple movies website were ready to start watching in about 15 seconds, while a 720p HD download of a game trailer from Xbox Live took 50 seconds to download the 108MB file.

But what about the neglected sibling of broadband speed: upstream?

The tests would seem to suggest I have between 1Mbps and 1.5Mbps - but what does this mean in the real world?
I'll be testing these upload speeds with some online gaming later in the week - and inviting you all to come and join me in a game of Call of Duty 4 or Halo 3 - but I started with a simple test.

I uploaded five high quality Jpegs - about 17MB in all - to Flickr and the process took 90 seconds to complete.

So that's Day One of testing. I welcome your thoughts and questions. You can follow some more of my tests and experiences on my Twitter feed @darrenwaters. I'll be sharing my gamertag with people on the Twitter feed later in the week.

Comments

  • Comment number 1.

    Interesting article - i had wondered just how much of the 50 MBps people actually received. I am looking forward to the gamer test results as that is the predominant use for my PC. More of a WoW man though... :)

  • Comment number 2.

    This is very interesting. I bookmarked - and will test my speed here in Machynlleth, Wales. Your five high quality Jpegs - about 17MB in all - to Flickr in 90 seconds sounds good.

  • Comment number 3.

    Gaming doesn't really say much about throughput... typically an online game will use less than 100Kbps of data in both directions.

    For gaming it is usually the latency of the connection between client and server that determines the quality of the experience.

  • Comment number 4.

    So what exactly is the point of paying extra for a 50meg connection? You might be able to download a file in 5 mins rather than 20 mins but you'd still have to leave it running and go away and make a cup of tea. And if you're going to have to do that then there is little advantage in having a fast connection.

  • Comment number 5.

    If I have read it correctly, this is what you will get when nobody else is active in your area. It would be interesting to know how many other 50Mb customets are in you area.

    Get them online and downloading and see how your connectivity behaves.

    Are you pulling a file from a locally hosted server?

    Gaming is not a bandwidth issue, you needs no more than 30kbps with a consistent 30ms response tme.

    I am lobbying for more transparency of the end to end service. Take a look at

  • Comment number 6.

    hi darren, Thanks for testing the 50 meg! you sound like your very aware of the web :-)

    i was thinking maybe for test 2 that you use a download accelerator like IDM (Internet Download Manager) and then pull down a file like ubuntu.

    also if you want another fast server to download from then apple quicktime is a fast server too,

    all the best,
    Spinifex.

  • Comment number 7.

    Completely agree with post #3 from p5yt3k - you need to go for something that will actually stress the upload bandwidth. Have you got a Slingbox? That'd make great use of the extra upload speed :-)

  • Comment number 8.

    Bittorrent and the like are famously greedy for bandwidth, download a few things and leave them shared for upload and see how it performs.

  • Comment number 9.

    As a user of Virgin's 20Mb service, it does deliver what it promises (outside the traffic shaping hours, of course!)
    At about 11:00 on Sat, downloading the latest release of SuSe via BitTorrent, (arguably the best test of any connection's throughput), the transfer rate was sustained at 20Mbps for almost 30 mins, so I would expect to download the same in ~12 mins on your XXL connection.

    Similarly, uploading via BitTorrent is also a very good indication of the connection's capacity, and would be just as effective a test on your 50Mb line.

    I will confess that I cannot remember if either of the speed tests on either Virgin or thinkbroadband.com use multiple connections to assess the speed, so you may see a more realistic performance estimate by saturating the connection

  • Comment number 10.

    Would be interested to see how well a HomePlug AV would cope with the bandwidth.

  • Comment number 11.

    I agree with Spinifex - a download manager will stress the connection more. I'm using Free Download Manager which downloads files in up to 10 (configurable) sections. I can often hit the maximum download speed with this.



    And to monitor network speeds I'm using the free NetMeter from:



    (URL looks odd but is correct)

    As well as showing instantaneous up/down and a short graphical history it gives you daily, weekly, monthly up/download totals too.

  • Comment number 12.

    I presume that Virgin's 50Mb service is still subject to traffic shaping? This seriously turned me off of their 20Mb service, and will continue to do the same to this even if it were priced reasonably.

 

The 大象传媒 is not responsible for the content of external internet sites

大象传媒.co.uk