What is this nancy boy nonsense?

Internet ‘rationing’ needed as UK cannot keep up with demand

What?

“The Internet is already consuming at least 8 per cent of Britain’s power output, equivalent to the output of three nuclear power stations, and demand is soaring,” Andrew Ellis, professor of optical communications at Aston University, told the Sunday Times.

He said: “It is growing so fast, currently at an exponential rate, that, in theory, it could be using all the UK power generation by 2035.

“We cannot make all that extra power, so we will have to restrict or reduce access, perhaps by metering consumers so they pay for what they use.”

So, who in Britain is getting free electricity? Anyone?

No.

Therefore it is already metered and restricted, isn’t it.

Idiot.

39 comments on “What is this nancy boy nonsense?

  1. I spy special pleading for more monitoring of individual internet use, and BT to get a new set of hands into everyone’s pockets.

  2. I’ve been watching and comparing the Tele and Daily Mail outputs the past few weeks (what, get a life? Never!) and the rule seems to be that whatever is in the DM is in the Tele 24 hours later. As with this article. I think the Tele newsroom is now so anaemic that they’re mostly just copying and paraphrasing Mail content.

    Anyway, it is of course complete claptrap.

    The other claim from this bunch of weirdies is that the internet tubes are full up of datastuff and so we can’t have any more bandwidth because, apparently they haven’t thought of laying another cable or something like that.

    Considering the DM article also quoted this guy saying it’s a choice of what you use your wind turbine energy for, we can safely assume that we’re in absolute barmcake territory with this one.

  3. He’s a professor…he’s not suppose to know how anything works in the real world.

  4. “We cannot make all that extra power…”

    Sure we can. Just not with renewable energy methods that don’t work when it’s cloudy or the wind isn’t blowing (or is blowing too hard).

  5. As PJ O’Roark once said, if you think healrh care is expensive now just wait until it’s free. Free thing tends to cost more for everyone.

  6. Tin-foil helmet time. What the professor really means is that information access must be restricted if the Agenda 21 blob is to triumph.

  7. More exponential rate nonsense. If something looks exponential, its just on the steepest part of a logistic curve. The rate of growth will fall.

  8. When I worked for a company that made fibre, the line was that all the limits were in the coupling electronics, not the actual string, so that once fibre was laid it was good for decades to come, only the ends need to be upgraded.

    So, y’know, bullshit.

  9. I’ve had a quick flick through a couple of his presentations and that article does seem to be garbling his message.

    Put simply, his argument is that we are reaching the physical limits of the capacity of the fiber that’s in the ground and the new tricks for increasing its capacity need increasing amounts of power both for initial transmission and regeneration. All fairly standard stuff.

    He doesn’t say this but we also know that laying fiber is really expensive so at some point either we cough up for the “free” internet or there will need to be some form of rationing of data speeds.

    The bit about running out of power is just some noise to get his research some attention as its mostly esoteric techy stuff for those at the leading edge of research.

  10. Yes, he should be fired simply for the cliched and absurd use of the “trending to infinity and DOOM!” Fallacy. I doubt he believes it – everyone knows we’ll all be 45 stone and dead of obesity by then.

  11. The point being is that installing infrastructure and pricing that in is what utility suppliers do as part of their business.

    This is like arguing that we’re reaching some gas supply limit because the pipes can’t carry any more gas. So put another pipe in the ground then.

  12. So what the guy is saying is, we can upgrade it, but it’s raping Gaia so we don’t want to. Uh huh.

  13. Oh, and another thing. The more petabytes there are flying about, the harder it is for the spooks to monitor it.

  14. Roue le Jour – “Oh, and another thing. The more petabytes there are flying about, the harder it is for the spooks to monitor it.”

    But is the traffic growing faster than Moore’s law? Because the spooks’ processing power is growing pretty fast. If it is growing faster than the traffic, they can search that data with greater ease. Not if otherwise.

    What this is ignoring is technical innovation. I assume the next step is to get rid of the cable. Especially the last few hundred feet to our homes. Some sort of wireless (that is, laser or microwave) transmission. So whatever problems with the fibre, how much longer do people expect us to be using it? A few decades? Sure. A hundred years? Ya think?

  15. Perhaps it’s just me being ratty this morning, but just because some heliographer has found the pound sign on his calculator, we’re all supposed to grovel in shock and awe?

  16. Maybe we could add a network of high power lasers through the streets to supplement the explosive batteries, catastrophically fragmenting flywheels and cubes of nuclear waste powering it all.

  17. Ian B – “Maybe we could add a network of high power lasers through the streets to supplement the explosive batteries, catastrophically fragmenting flywheels and cubes of nuclear waste powering it all.”

    Attached to the heads of sharks? We have suffered from a lack of vision since the future ceased to be about flying cars. But I think this is a future that we can all unite around!

  18. SMfS

    Good question. I guess in a hundred years we will have turned the entirety of the em spectrum over to “internet” but that’s just what the old people will call it. Young people won’t have a name for it, like fish having no word for water.

  19. Roue le Jour – “Young people won’t have a name for it, like fish having no word for water.”

    Sure. Not much use for English either. All they will be able to say is “Resistance is futile, you will be assimilated.”

    They would say “Your biological and technological distinctiveness will be added to our own.” But they would be too afraid of Ironman screaming like a baby if they suggested anyone was biologically distinct from anyone else.

  20. The prof sounds about as switched-on as the noble lord who opined early in the 20th century that the motor car would never be a success as there would not be enough chauffeurs!

  21. To any techies on the thread – are we reaching the limits of DWDM to increase the capacity of the fibres?

    I’d have thought that improving laser tech, narrower bands, and higher bit rates meant that physical fiber isn’t that much of a constraint?

  22. @johnnydub
    Something I read recently, they were talking about fiddling with the laser light. Phase modulation or polarisation or something. Talking about data carrying capability being boosted by several orders of magnitude over the same fibre.

  23. Ukliberty,

    You might get your bytes at a flat rate but the baud rate is is what they base their flat rates on. And you don’t get your ‘leccy on a flat rate?

    /double pendant?

  24. Johhnydub:

    “are we reaching the limits of DWDM to increase the capacity of the fibres?”

    Short answer is no. Longer answer is that we might need to light up some more fibres on some of the core routes and even lay down some more cables here and there. Nothing terribly out of the ordinary though. The main constraint against this happening is if the governments continue to mess around with the price mechanisms through net neutrality rules and similar silliness

  25. The rise in power consumption is not from browsers of the internet, but rather the massive data server farms that are chugging away 24/7 to make that data available. This “internet” power usage will grow regardless of how much people are using the cables and infrastructure to get the data to them, it is the providers of that data that need power, and they need it whether anyone browses it or not.

    In fact as we move to mobile browsing the “user” power consumption has decreased per unit, sales of wattage hungry personal computers are decidedly dropping in favour of super efficient tablets, whereas those of even more wattage hungry data servers are increasing, together with their redundancy and the massive disk arrays where all the cat videos are stored, often in triplicate, and the air conditioning that keeps it all cool, all running all day every day.

    But the biggest flaw in the argument is that as power consumption moves to internet usage it moves away from other vastly more power hungry consumption. Ordering
    online and having it share a single van is much better than a few dozen individual car journeys to the shop or supermarket. Working from home doesn’t need passenger energy costs. Reading online news avoids all that energy required to print and transport dead trees around, which if they stay alive instead are negative carbon emitters.

  26. @smfs

    “What this is ignoring is technical innovation. I assume the next step is to get rid of the cable. Especially the last few hundred feet to our homes. Some sort of wireless (that is, laser or microwave) transmission. So whatever problems with the fibre, how much longer do people expect us to be using it? A few decades? Sure. A hundred years? Ya think?”

    If anything we are going more to fibre as there isn’t the bandwidth in the radio spectrum for core uses and the more we push out broadband technologies to the user like 4G the further fibre has to go in to the network.

    PCCW had a massive fibre project in HK a few years ago as it was the only way they could support their VoD and public WiFi.

    Having worked on fixed wireless to the home its a technology that is just too expensive and has been for 40 years. We even looked at using lamp posts and got a trial working in Reading but we couldn’t get the business plan to work. It doesn’t take many users all wanting VoD for you to need a big fiber connection to the base station, at which point you may as well have fibre to the cabinet and use copper for the last mile.

  27. The even longer answer is this:

    a) with current technologies we are basically putting one beam of light down one strand of fibre. This is called ‘single-mode’. It is possible to construct fibres which have more than one optical core per strand. This is known as ‘multi-mode’. If you have twenty cores you can in principle put twenty times the data down a given fibre. Multi-mode fibre has the drawback that it is more expensive than single-mode and has higher attenuation, so it needs amplification and regeneration at more frequent intervals. But progress is being made constantly. By the time Prof. Whatsit’s putative Eight Years To Save The Internet rolls round, people will have been working for eight additional years, and the state of the art will be that much further on.

    b) rather conveniently, fibre made from silica (glass) has a low attenuation window in the 1500nm (near infra-red) about 50 THz wide and a low dispersion window around 1300nm. The ~1500nm band is where most fibre is used. The Shannon-Hartley theorem says that even with a 1:1 signal-to-noise ratio you can stuff 50 terabits/sec down this, and with 30 dB SNR it’s more like 500 Tbit/s. You can increase SNR by making the signal larger or by decreasing noise. And if you keep your transmission rate below the channel capacity then you can get error rates arbitrarily small.

    c) increasing raw data rates onto the fibre is no good if the transceivers can’t handle it. Increasing the operating speed of optoelectronics is a huge and deeply-funded R&D effort, and once again strides are being made every day. Any hint that progress is slipping will mean investment automatically swings towards areas that need attention.

    d) the dominant cost in laying fibre is not the cable. It is digging the street up, installing the ducts and making good afterwards. If the initial work is done right, then laying additional fibre costs a fraction of the original. Upgrading the optoelectronics is expensive, but it’s not a big bang sort of thing but just something that happens as a part of the normal tech lifecycle.

    e) the energy cost of providing telecomms services is a complete red herring. It can be priced in just as it is into any other energy-consuming economic activity like smelting aluminium or running the dryers in a laundrette.

    The market will sort this out. It’s as good an example of what markets do best as I can think of. As anyone who’s done IT provisioning knows, Parkinson’s Law imposes costs. My prediction: there will be no bandwidth crunch, and if anyone is suggesting otherwise he’s probably out for a buck.

  28. We were doing frequency division multiplexing on fiber 25 years ago.

    What should frequency division of light mean to you? COLOR. Multiple transmissions over the same strand using different colors (!). When I was a network manager in the late 80s.

  29. Something I read recently, they were talking about fiddling with the laser light. Phase modulation or polarisation or something. Talking about data carrying capability being boosted by several orders of magnitude over the same fibre.

    They were talking about firing “tuned lasers” down fibre optic cables for broadband when I worked for Marconi 15 years ago.

  30. “Tuned lasers” is DWDM – i.e. Dense wavelength division multiplexing.

    It was more a question as to whether we wre encountering limits on the number of parallel wavelengths.

    I sold that kind of gear 10 years ago, but now I’m selling consultancy in a different area of the IT industry, hence my question about the limitations.

    From the answers above, we don’t seem to have a problem; sounds like this professor is rent seeking.

  31. Bloke in Costa Rica, that is not what single mode and multi mode means. Multimode fibre is a bigger waveguide (50-100 micrometres as opposed to 8-10 micrometers). It was more expensive to produce the fibre (more optical glass, way more) but the equipment to transmit on it was cheaper until about 10 years ago because you could use diodes rather than lasers. This is no longer true. The range and bandwidth you could get out of it was less because of modal dispersion (some light waves take a longer path, some shorter, because they have more room to bounce around), smearing out the signal. No one in their right mind would install multimode fibre anymore.

    What you’re talking about is multiple single mode waveguides embedded in the one substrate. Which means, hey presto, replace all the fibre! This is the problem with people who claim that once you install fibre, it’s forever and you just have to upgrade the end points. What if the new technology requires a different grade or design? Oops. It’s the pits and pipes that matter, i.e. how easy it is to replace.

    Anyway, the eminent professor was talking about power usage, not bandwidth. I’m very dubious about his predictions.

  32. @Johnnydub,

    We don’t appear to be running out in a practical sense but this guy’s looking ahead and someone has to be at the bleeding edge. Its normally driven by manufacturers, as you know, who will have probably shifted as many DWDM boxes as they are likely to and need to start working on the next technology in order to get the standards and patents sorted by the time its needed. Or more cynically in time for the salesmen to start hypeing a capacity crunch 🙂

    Despite 4G roll out being nowhere near completed I’m already getting questions about whether the sites we having built for rural coverage will support 5G. To which the answer is it hasn’t been standardised but the vendors are already working on it for when they stop shifting large volumes of 4G boxes.

  33. Ltw: yes, you’re correct. I misspoke. Should’ve said “multi-core” fibre.

  34. BiCR – fair enough. It’s an interesting technology, but I’m a bit dubious as to how it would work in the field. Stripping, splicing, and terminating single core fibres in the back of a truck or in a tent is dodgy enough already. Aligning multiple cores for a fusion splice or designing reasonably robust and reliable connectors for the patch leads sounds like a fucking nightmare 🙂

Leave a Reply

Name and email are required. Your email address will not be published.