4 ms·

OK then, fair enough. Here is me talking about his scientific illiteracy at least four years ago to give people a rough idea: https://delimiter.com.au/2014/06/

• GrahameLynch 8y ago
  Rod Tucker originally modelled a wireless only NBN as having access to only 20MHz of spectrum and that this represented the sum total of what the nation's broadband demand would be funnelled through if a policy decision was made to use wireless in the last mile. Do you regard Tucker's 20MHz premise as scientifically accurate or plausible given the NBN actually does have around 75MHZ of spectrum that it uses to serve just 3% of its base with wireless? I take it you would endorse that as the counter-factual. The fact is that there is scores of gigahertz available for the communications needs of end users. The V-band alone (which is being used for last mile for various applications) has 35GHz of spectrum alone. https://en.wikipedia.org/wiki/V_band https://en.wikipedia.org/wiki/V_band mmWave bands add another 14GHz of spectrum. https://en.wikipedia.org/wiki/Extremely_high_frequency https://en.wikipedia.org/wiki/Extremely_high_frequency Four years ago I said there was 80Ghz available in service of the nation's communications needs. You interpreted that to mean the needs of a "government owned last mile fibre network". The point I was making is that it is fraudulent to model the limitations of a technology (in this case, wireless) on the basis that any given network only has the use of a tiny fraction of what it could be using (and indeed does). Sorry everyone for hijacking this thread but hopefully it means that the quoted points I made above re China and WCIT make more sense, because I actually do have more than half a clue.
  • quink 8y ago
    > Rod Tucker originally modelled a wireless only NBN as having access to only 20MHz of spectrum and that this represented the sum total of what the nation's broadband demand would be funnelled through if a policy decision was made to use wireless in the last mile. Yes, but he also accompanied that with saying broad coverage would imply a tower outside every 2nd premise. If only there was someone who would know with absolute certainty that Rod Tucker did imply that. Cue the very informed guy who wrote this: http://www.couttscommunications.com/Published-Articles/cd140613commsday.pdf http://www.couttscommunications.com/Published-Articles/cd140... And while it's not a lot, 20 MHz is an OK starting point if you have a point-to-point service doing a basic service to a fairly small percentage of premises. He did the modelling in 2010 and NBN didn't buy Austar spectrum until 2011, which at the time was allocated already so far from predictable that the government could find any spectrum for the NBN before the digital dividend, so yeah they have more than 20 MHz now, but before then who knows. > The point I was making is that it is fraudulent to model the limitations of a technology (in this case, wireless) on the basis that any given network only has the use of a tiny fraction of what it could be using (and indeed does). Was millimeter wave even remotely ready for viable commercial rollouts for last-mile access in 2007 to 2010 when this discussion was meant to have been happening? Hell, it's not even a thing now about a decade later, so what were Australians to do in the meantime? Also, is it viable to limit model the limitations of a technology if it's in wide use? Because I know this guy who was quoted by Turnbull himself https://www.malcolmturnbull.com.au/media/coalition-broadband-policy-frequently-asked-questions https://www.malcolmturnbull.com.au/media/coalition-broadband... I think is name is "Tucker, R" who said in this paper: https://people.eng.unimelb.edu.au/rtucker/publications/files/energy-wired-wireless.pdf https://people.eng.unimelb.edu.au/rtucker/publications/files... cited 169 times apparently that "The power consumption curves for DSL, HFC, and FTTN cease prior to 2020 because we believe these technologies have a limited ability to scale and meet future increases in bandwidth requirements.". So is this the only seemingly arbitrary limitation on being allowed to model something or is HFC still the best thing since sliced bread? Edit: Also of note, the picture of "… and another 2010 slide showing a DSL node outside every second premise" is actually utterly remarkable in how exactly like FTTdp, which is apparently the latest saviour of Malcolm Turnbull's Mess, it is about a decade prior to it actually becoming a reality except it's in the pits rather than on the strips.