Kevin Rudd’s announcement that the government will build a network that delivers a 100Mbit/s fibre to the home Internet service, within 8 years, has lead to a flurry of analysis in relation to commercial viablity. However I have not seen much discussion about the technical aspects of the proposal. To be sure a 100Mbit/s home Internet connection would be a fast connection today. The top of the range home Internet service is typically around 24Mbit/s at the moment and you pay a bit for it. Usually such a service is delivered via ADSL2+ technology using traditional copper pairs. Services with a similar speed are also available via Hybrid Fibre Coaxial (HFC ) from Optus or Telstra (which is also the delivery medium for most Pay TV services). Relative to these existing services 100Mbit/s looks fast. Over 4 times faster in fact. However Kevin Rudd’s broadband network is not a service being delivered to the market today. This is a service that if built may not be available at your home for up to 8 years. So the question is will 100Mbit/s still be fast in 2017? I very strongly suspect that it won’t be. Here is why.
Firstly lets look at the pace of change. Network speeds are today increasing in much the same way as computer processor speeds have been increasing for the past 40 years. In 8 years time it is not unreasonable to expect network speeds to have double four times over. Or in other words it is not unreasonable to expect network speeds to be 16 times faster than they are today. Under such a scenerio that would mean that the top of the range 24Mbit/s service will be replaced by a 384Mbit/s service. Just as the top of the range home Internet service 8 years ago (in 2001) was 16 times slower than it is today at 1.5Mbit/s. And 8 years prior to that (1993) we would have all been thrilled if we could have had 94kbit/s although only if we had know what to do with it.
Some will counter that ADSL services are surely reaching the technical limits of copper. However there is no reason to believe this is the case. Just 18 months ago the University of Melbourne caused a global sensation with news that one of it’s research students had figured out how to build an ADSL service that delivers 100Mbit/s. And 2 years ago others were suggesting that the theortical limit of copper telephone wires might be around 100,000 Mbit/s. That suggests that there is a lot more juice to be squeezed from the copper network. And none of that entails digging up my drive way.
What about the HFC networks? Well we know they can go a lot faster because a month ago Telstra announce that HFC home Internet services in Melbourne would all be increased to 100Mbit/s. And this won’t even be a small fraction of the capacity that can ultimately be squeezed from this medium.
However fibre can run a lot faster than 100Mbit/s I hear you say. Well yes it can. Which kind of makes the point that what the government has announced is actually the construction of a rather slow network.
For all the talk from the neo-socialists about privatisation of telecommunications been a bad thing because it leads to duplication of infrastructure, all of a sudden a government plan that delivers a slow network via a huge duplication of infrastructure is somehow a good use of taxpayers money. For all the talk about private broadband services being too expensive suddenly a service that is going to be even more expensive is better because it’s government built. Will the wonders of neo-socialist spin never cease?