there has been a bit of coverage over the last few days after virgin media in the uk announced a trial of docsis 3.0 cable running at 200mb. of particular interest was the claim that “the 200mb pilot is believed to be the fastest implementation of docsis3 technology in the world, running faster than services offered in japan and the us, which currently reach 160mb and 101mb respectively”.
what’s even more interesting is the subsequent rebuttals of this claim by other carriers and a flurry of articles debunking the claim that cable can actually deliver real-world performance coming even close to speeds like this. this is based, of course, on the fact that cable is based on a shared access architecture, with any number of subscribers sharing not only the backhaul from a cmts, but the physical cable running into the premises as well, in a giant daisy-chain. this is distinct from dsl, where each subscriber is connected by a dedicated, physical piece of copper. this may not run at the same headline-grabbing speed (adsl2+ in new zealand might demonstrate real world averages at 12-15mb; vdsl2 maybe twice that), but it isn’t shared with anyone else. and backhaul from a dslam is typically gige - 5 times as much bandwidth as on a shared 200m cable segment. fibre to the home, in the form of gpon, shares bandwidth between a number of homes in a similar fashion to cable. but where cable may run up to 200mb, gpon is sharing 2.4gb between 64 or 32 households. a fair step up.
either way you look at it, it’s an interesting discussion (well, argument). personally, i think vdsl2 is a good way to go, particularly here where there is little cable infrastructure (outside of areas in wellington and christchurch serviced by telstraclear), and adsl2+ can provide a solid migration path. and thats without considering bonding or vectored xdsl, which may potentially provide additional significant speed boosts in the future.
- BROWSE / IN TIMELINE
- « 2degrees
- BROWSE / IN Uncategorized
- « 2degrees
