|
Ethernet is irrelevant though to the point of this discussion. Try pumping 100Gb/s over ethernet cables of a couple of hundred meters in length and see what happens.
It is also not really valid to compare current DSL speeds to dialup. The basis of the technology is different. Dial up was cramming as much bandwith as possible into a voice architecture. If you ever wondered why 56K was the limit for a dial up modem, it was not really the case, 64K would have been, but we had to follow the bigger market so we only got 56K. A voice channel on the POTS network is 8 bits at 8Khz giving a 64Kbps data rate, that would be the max for a dial up modem and is limited by the technology placed on the ends of the copper pair, not by the pair itself (In a perfect world, obviously lots of people had copper in poor enough condition to prevent seeing connection speeds as fast as 56K) the reason we got 56K not 64 is that the larger markets used 8 bits at 7khz for a voice channel rather than the 8 bits at 8khz we do which gets you 56kbps, not 64 (I may have the figures the wrong way round there and 7 bits at 8khz, but you get the picture)
From there we have moved on from the termination equipment being the limiting factor to the line being the limiting factor. They can't just magic away the fact that a couple of hundred to a couple of thousand meters of copper pair has limited bandwidth. It is about maxed for ADSL flavours and VDSL is going to be right on the bleeding edge of it's abilities to deliver the Libs plan to a significant portion of the population assuming good quality copper. Most of the other magic bullets proposed by the liberals over the last couple of years are pair bonding variants where they use multiple pairs for simultaneous transmission that is aggregated at the ends. Given how many housholds struggle to get one pair in good enough condition and suitable architecture for DSL, how are they going to get multiple pairs?
You also can not use the increases from up to 1500K to 8K to up to 24K connection speeds to say how much copper technology has improved as they have been driven by commercial considerations as much as technical ones. The last in particular relating to the switch on of ADSL2+ which Telstra has a history of not doing in any given exchange until a competitor installs a DSLAM there and enables it first even though the equipment in the exchange may already support it. The original 1.5K to 8K "Increase" was the dropping of imposed limits by Telstra rather than improvement in DSL technology.
The fact remains that even if built on time and on budget an FTTN plan is going to be struggling to meet demand by the time it is finished and will by design, impose limitations that really do make it into an entertainment delivery medium with content producers at one and and passive consumers at the other. It is like setting a 40KMH speed limit on the hume freeway with a speed camera every 500M to ensure it is adhered to and saying that you obviously do not need a higher limit as everyone is doing 40. Except that in this case there would have to be a dedicated lane for a lucky few who have a 110KMH limit and theothers will be going at varying speeds depending on how lucky they were.
Given that it will be struggling to meet demand by the time it is finished, why waste billions of dollars on an obsolescent network that will be up for replacement in a fraction of the timeframe that its initially somewhat more expensive, vastly technically superior, already available and already started alternative would be?
I have thought pretty long and hard (As I work in the industry) about the design complexities of FTTN and and from a pure design perspective it is more difficult to achieve than FTTH. FTTH you design from scratch and put the network where you need it. FTTN you are constrained by the existing network, you need to hit a certain number of homes per node with a maximum distance between node and home in the nhundreds of meters and have to examine the existing copper network, which was built on the assumption of voice services which could be run over tens of kilometers without major problems and try to find somewhere to put nodes that will both service a reasonable number of homes AND not be too far away from any of them.
I can see significant amounts of copper network having to be hauled out and replaced with new copper in our future either to remedy poor quality or insufficient cable capacity.
|