On Mon, 02 Jan 2012 17:14:18 +0000, Andrew Benham churned:
>
> The cable cabinets cover fewer premises than the OpenReach PCP cabinets,
> so you're likely to be closer to a cable cabinet. The attenuation on
> coax is less than on twisted pair. These two combined mean you're
> likely to have less loss on a shorter run of coax than on twisted pair,
> so you win twice.
> I think Virgin Media claim that their customers do better in the "up to"
> speeds than customers on twisted pair ADSL. Some figures mentioned on
> http://www.broadbandchoices.co.uk/tools/speed-test.html
But their network infrastructure is different. Ultimately *everything*
they sell ends up going back to a regional centre in a single, typically
oversubscribed, pipe.
If the VM network was so wonderful [which it is not, it is shit], I'm
sure the need to saturate your data centre with SQUID type proxy servers
would be completely unnecessary.
FTTC has some similarities to cable - they both seem to be victims of
more than normal (if you can define such a thing) amounts of packet loss
and delay. My money, however, would always be with FTTC over cable any
day of the week. Not only do you have a choice of provider, but even on a
bad day BT still have far better infrastructure (even the 20cn ATM back
haul) than VM will ever be able to afford to invest in. Jesus, just look
around any housing estate at the state of VM's shit - it's an utter mess.
If that's the bit they have on show to the world, imagine what the rest
of it is like.
That ole 'speed' metric is utter bollocks. I may deliver 75% of your
packets 10% faster than your competitor. But if I drop 25% of your
packets forcing retransmission a few times will you benefit that much?
The whole gig is about more than the local link speed, it's the back end
throughput that really matters.