[Ignitionnet]

Originally Posted by weesteev

The upstream launch with Broadband Size XXL of only 1.5Mb is nothing to do with core capactiy but more to do with the "pre cert" headend gear. The initial trial of upstream channel binding was done with Arris CMTS equipment which is silver cdertified DOCSIS 3 and allows upstream channel bonding. I believe 4.5Mbps upload was tested at this phase by bonding three channels together.

As I remember 5 or 6Mbit was tested. It's also no excuse, having seen 5 and 8Mbit offered on unbonded upstreams.

In any case as was mentioned here apparently VM are only using a single upstream for the 10k initially due to being somewhat spectrally challenged in areas so bonding is a non-issue.

The equipment that has passed standard is the Cisco 10k series uBR chassis, both Docsis and Eurodocsis standards are being trialled currently and wil no doubt be franchise specific upon completion.

It'll all be EuroDOCSIS. Virtually all modems on 20Mbit will be dual-mode so would be a waste of bandwidth to run with DOCSIS. Modems that aren't EuroDOCSIS just won't end up on the 10k.

The modem that is being supplied will be the rather sexy lookig Ambit VMNG 300 Docsis 3 modem. It is very different to current modems and is a matte polished finish to bring it inline with more modern equipment (think playstation 3).

So long as it works that's all good, though not sure about calling it the 'VMNG', PR even on the modem, bleh.

Virgins total speed is calculated on all three broadband tiers (2mb, 10mb and 20mb). Considering that 2.4million of its customers use 2Mb, thats quite an impressive result to have!

Without knowing the breakdown of customers using the speedtests it's impossible to say how impressive or otherwise that is. Going by another website quoted by VM in the past the average speed on 20Mbit is just over 6Mbit.

Faster speeds have inherent problems for all networks, how many non PC enthusiasts have high speed broadband from "legacy" cable days and may not have the PC (or modem) to support such a fast speed.

They should have modems that run just fine for speeds, and a speedtest really doesn't stress any PC to run, a hard drive that can't write at 2.4MB/s is in real strife. This will also affect all operators equally.

The speed tests dont take how the testers connect to the internet, the avergae wireless user will be "G" based, severley restricting most speed test results.

As will this

So overall, speed test sites are rubbish! They will never give you an varying result of your speed! If you want to judge your speed then use a newsgroup server and download large files to judge your true speed.

Discuss ?

Newsgroups will give an even worse result on an old PC due to the stress of having to decode the stuff coming down. Wireless restricts newsgroups equally.

The only way to really test is to connect directly to a modem with a shiny new PC and have a dedicated link.

I am not really that surprised that VM might score well in a speedtest, however one would question whether it's something to be proud of when you're the only ISP in the country that doesn't have distance limitations on the service they can provide, and are only on average delivering your top tier of customer 36% of their line rate, about 1% higher than your copper based competition. Hyping yourself and your fibre optic broadband to death and being matched and in some cases beaten by the DSL crowd isn't good.

I wonder how much higher the speed test results would have been without STM in place