BBC BLOGS - Newsnight: Susan Watts
« Previous | Main | Next »

Two go mad in Silicon Valley...

Susan Watts | 16:35 UK time, Wednesday, 29 October 2008

Our trip to California turned out to be quite nostalgic. For me, it was a throw back to my days as a cub reporter on Computer Weekly in the late 1980s, when we still banged out our copy on typewriters. For Newsnight producer, Ming Tsang, it was a mini refresher course in the history of computing. Read his thoughts here.

We set out to make a film about the huge amount of energy used by the giant computers that drive the Internet, and our data-obsessed lives. Everything we do which requires tapping into a keyboard, uses one of these computers, called servers. They're kept in enormous climate-controlled warehouses known as data centres, or "server farms" to geeks and governments. And these are getting so big, and using so much energy, that people are starting to take note.

In fact, for every unit of energy a server uses to actually do some computing, it takes an equivalent amount of energy to cool it down again. US Government statistics show that data centres now use as much energy as the whole of the car manufacturing industry. Given that the US alone builds around 10 million cars a year, that's an awful lot of energy. And it looks even worse if you stare into the future.

The latest industry figures suggest that by 2020, the carbon emissions produced in generating energy for the Internet will be the equivalent of those produced by the airline industry. These figures assume that neither business makes much progress in cutting emissions, so we set out to ask some of the computer industry's leading players how the industry plans to use less energy, or to switch to cleaner, less polluting sources.

We met Three Wise Men - or at least Three Green Tsars - each at a computer company that represents a grand era of computing. In their time, each of these companies has been, if not the biggest, then the most influential on the scene. First IBM, which created a commercial market in supercomputing in the 1950s and 60s. We were invited to the company's Silicon Valley research labs, to see the biggest data centre in the world. Well, it was when it was built, in 1980. Nowadays, it's just average-sized. IBM's wise man told us about the company's plans for virtualisation and "cloud computing" to cut the energy it uses.

Next stop, Cisco, just down the road. In the 1990s, it was Cisco that built the routers that run the Internet. The Green Tsar here took us into one of the company's server development labs - a complete contrast to the neat, tidy functionality of the IBM server room. Here we saw in action the research that's creating coming generations of the Internet - web 2.0 and beyond.

And despite the best efforts by Cisco's green team to use tricks that cut the energy used by the Internet, it persists in growing. After all, a few years ago there was no BBC iPlayer or YouTube. Now, both of these video-on-demand services need powerful, energy-hungry servers to stay on air.

So, to see how Internet companies view the problem, we met our third company, the owner of YouTube, Google. If you're reading this online, I assume you've heard of Google. Not only is Google rich and powerful, but it has a stated aim of helping to protect the environment. Oh, and it's famous too as a cool place to work. So it was with some anticipation that we approached Google's mission control, "Googleplex".

Google's Green Tsar took us up onto the solar rooftop of its Mountain View headquarters. Strange to think that under our feet could well be some of the rooms full of servers that Google's PR team didn't want us to film - so closely guarded are the secrets in those rooms, as are the locations. The secrecy is aimed partly at protecting data that belongs to Google's customers', and partly at protecting Google itself...

As for the environment, we'd heard all about Google's plan to create "energy cheaper than coal". This is apparently to be a gift to the world, as well as helping to ease Google's corporate guilt over the carbon emissions it creates as it uses more and more energy to drive ever bigger data centres.

But apart from a few minor investments in alternative power sources, Google couldn't tell us a great deal about its approach to cutting energy consumption, and definitely wouldn't tell us how much energy it uses at the moment. It's a lot. Think how many searches you've carried out, and how many video clips you've watched on the net today. And with its latest mission statement to "organize the world's information and make it universally accessible and useful", this can only grow. Yet Google seemed surprisingly reluctant to spell out to us how it plans to polish up its Green credentials.

Finally, we heard from the "Green Grid" industry body set up by computer manufacturers to curb the Internet's energy appetite. But in spite of the name, Green Grid wasn't formed wholly out of eco-concern. Selling more boxes and avoiding government regulation had something to do with it too.



  • Comment number 1.


    As Susan said nothing of the (green) algae initiative, when in Hawaii, I wonder what has been missed in Silicon Valley?

  • Comment number 2.

    Is it not strange that Newsnight asks questions that it was not interested in having the answers to in 2001.

    The favourite proposal in the competition for the Dome was the global environmental centre.

    This was the one that should have won on the criteria, it was the one the Government's own consultants wanted to back.

    It was the one where the finance was lined up if it was successful.

    Newsnight and the BBC had no interest. I assume because it was about environment and science and not mainstream politics.

    So the media allowed the Government to 'force' the casino/venue on us.

    Why ask questions now, when you didn't want the answers 7 years ago.

    £400 million was allocated in the start up alone to the computing, processing and systems monitoring alone.

  • Comment number 3.

    More climate change nonsense and, frankly, hysteria based on poor science.

    I do not disagree with the likelihood of human activity being a contributory factor in climate change but I do object to every possible example of energy usage being pushed down our throats as an instance of a problem that will inevitably get worse.

    Jeremy Clarkson is forever telling us that the computing power in a 5-series BMW is greater than that which sent man to the Moon. The energy consumed by the engine of this car to run the computing is a tiny fraction of that required to run the Apollo mission hardware.

    The phone I have just got will surf the web, send emails, play video and audio, take and store photos, run a GPS navigator - and make voice calls - with a battery of only 1500 mA/hrs. And it will do all this for days at a time on one charge.

    The so-called problem that Newsnight is scratching at is that the demand for the service is being placed on an as yet rather primitive form of delivery.

    There is no energy crisis in the delivery of computer services such as the web and point-of-sale financial services. There is a temporary peak in the demand/energy-required equation that will be resolved in a few short years. The examples I have given - and there are zillions more to choose from - illustrate this clearly.

    The BBC itself has run a piece on "cloud computing" in the past day or two. It's clear that within a few years, all these lumps of hardware, many millions upon millions of them sitting in offices, shops, factories and homes, all consuming 300 to 500 watts each [and in the case of many commercial premises and homes, left running 24/7] , will be replaced by a connection from a keyboard to a server somewhere that provides the computing and internet access that we all now cannot do without.

    Newsnight is the flagship of BBC current affairs reporting, we are frequently told. This one is a none-story that has enabled your reporter to have a jolly in sunny California., and far from ‘flagship’ standard material.

    By the way, what is ‘an advance preview’, which is what Watts’s blag is posted as? Please can she arrange for me to have an ‘advance preview’ of next Saturday’s winning Lotto numbers? Or is that ‘an advance prior preview’?

    Yours truly, Chris Nation. [Unsuitable/Broken URL removed by Moderator]

  • Comment number 4.


    Anyone see the connection with Ross/Brand?

    Do the BBC seek the 'younger audience' in the hope they will not notice the decline?

    (Please imagine this is emailed 'on the tilt', with a blurred black-and-white shot of my feet, taken from a helicopter, through selected obscuring clutter, accompanied by din-drowned commentary. I would hate to appear un-cool.)

  • Comment number 5.

    Why do I have to watch Newsnight to see a scientific programme that should be on prime time TV? I refer to Two go mad in Silicon Valley. Now we know that computer servers are big consumers of power the next step is to draw up a comprehensive list of major industries in descending order of power consumed. I suspect that the entire global radio and tv transmission and reception resources, including military, aviation, navigation, radar and anything else you can think of might be high on the list, perhaps with all the resources involved in generating the transmitted signal included. It was ironic to see pictures of all those servers with their cooling fans adding yet more power. In the beginning it was the valve with its power hungry heater, then the semiconductor, the next stage must be a device that consumes no electricity, optical for example. Thanks for an excellent programme.

  • Comment number 6.

    I understood your film to be about internet servers and the data centres they inhabit, so my comments relate to that.

    I would imagine most of the heat produced at these places by servers , is from the data storage devices, disk drives to you and me.
    Power saving on Servers is rarely enabled , mostly because of the 3 -5 second delay hard drives take to power up(spin up time).
    What I feel you failed to take into account was the newer type of Data Storage Devices now hitting the market place , Solid State Disks. Not only do these require less power and produce less heat but are a lot quicker in Access Time (sector skipping). As with any Server,performance is the No.1 priority.
    Also with no moving parts there is no disk spin up delays , servers will be able to use ACPI like power saving for Data Storage Devices for the first time , Solid State Disks can be powered up or down in milliseconds , unlike the older hard drives that had 3 -5 seconds delay.

    On a different note , when mainframes were water cooled some places used the hot water for heating other offices , there are signs this is coming back into vogue , look at some of the newer mainframes and they have the option of being water cooled again.

    Just my take on the film.

  • Comment number 7.

    Last week I read in our local newspaper that a data centre is being planned for Caithness, to be powered (and cooled) by cheap power from tidal energy in the Pentland Firth.
    The other suggestion was that the surplus heat could be used to heat greenhouses growing vegetables, etc.
    Let's keep up the positive approach to the technical problems facing us.

  • Comment number 8.

    Slightly O/T, but related and as there is qualified brainpower here that might have access to objective commentary and/or supporting links...

    Yesterday I had another insight into the confusion on green issues that we all face.

    En route to a meeting, I listened to the BBC radio news, with a story about our seabird population.

    Not good. Seems they are in a dire state, and this can be traced back to the state of their principal food source, sand eels. And this... is due to global warming.

    Oh dear.

    Thing is, just 15 minutes later the DJ, Terry Wogan (somewhat of a rebellious spirit to the corporate agenda, like they don't have enough already) read out an email on this story (though without any reference to the news slot just read out) from a listener of unknown professional provenance, that the sand eel reductions were nothing to do with climate, but more at the door of Danish fleets 'hoovering them up' (the term used) for animal feed and even, ironically, to be used as an alternative source of fuel.

    Now, as a listener, in one quarter of an hour I had two diametrically opposite reasons provided for the fact of sand eel reductions over the airwaves and am currently none the wiser as to which is anything like the accurate one.

    Now, I hate to say it, but there has to be a problem that when it comes to reading out any old press releases by charities that cite 'global warming', the track record of this £3.5B, multi-tens of thousand strong (analysts, researchers, etc) national broadcast entity is a bit suspect, and the net result of all this has been to cast more doubt on anything the BBC comes out with on science issues.

    Any insights?

  • Comment number 9.

    It's easy to cool a data centre. Build it next to a heat-only energy from waste plant, and use the spare heat to run absorption chillers. You win out several ways then, because you can :-

    a.) Offset the impact of landfilling that residual (you should really recycle 70% plus of your waste - EfW is a residual treatment technology.)
    b.) Offset the cost of generating electricity at 35% efficiency in fossil-fuel plants to inefficiently run refrigeration plant to cool these chillers (it works just as well on heat.)
    c.) Ensures that the efficiency of the EfW plant is as high as possible (you can get over 70% efficiency on a 'heat-only' plant. Traditional EfW for electricity generation doesn't get to 20% net efficiencies. Waste isn't a terribly good fuel if you want to generate electricity due to its lowish calorific value.)
    d.) Provide a year-round heat-sink for this EfW plant to use, without which you can't really run a 'heat-only' solution (even district heating tends to be very patchily used in the summer - conversely, this would be your highest demand for heat if you used absorption chillers.)

    I've looked at the lifecycle assessment of this and the CO2 balances, and it's quite an elegant solution. The engineering is quite simple too.

  • Comment number 10.

    Hello Junkkmale,

    The lack of Sand Eels was originally blamed on fishing in the area so no-fishing zones were declared. However the decline in Sand Eels persisted.

    The factor that best explains the movement of Sand Eels is ocean warming. And the substantial global heat imbalance implied by the degree of ocean warming measured since the 1950s can only be explained by a global energy imbalance between what's coming from the sun and what's going out to space.

    This energy imbalance is best explained by the Enhanced Greenhouse Effect caused by human CO2 emissions. i.e. Global Warming.

    The moral of the story - there are masses of people out there with opinions not supported by the evidence and access to email. But they lack the background knowledge of the experts who you'll often see the BBC quote.

  • Comment number 11.

    Thanks CobblyWorlds,

    Do you by chance have a link to the findings of the experts the BBC quotes in this regard.

    Not the Terry Wogan bit of the BBC... the other one.

  • Comment number 12.


    I don't know who the BBC quotes as I don't know what the article is, but here's information from someone other than "advocates" like environmentalists:

    Fisheries Research Services: Sand Eels
    Discusses the measures taken with regards fishing and their lack of impact (particularly last 2 paras).

    "Regional and annual variation in black-legged kittiwake breeding productivity is related to sea surface temperature." Fredericksen et al 2007.
    Which finds "A negative relationship between breeding productivity and winter SST in the previous year was found within 2 regions (East Scotland and Orkney), as well as in a cross-regional analysis." SST is Sea Surface Temperature, "negative relationship" means "SST goes up, productivity goes down".

    If fishing had caused the drop in Sand Eels in the Western Isles, controlling it should have increased population. That it did not, while another key factor continued (SST increases), strongly suggests the blame lies with SST. Not the local fishermen. That is in this Scottish case; other fisheries may be subject to different factors.

  • Comment number 13.

    Hello Susan,

    A voice from the past...I remember the day that cub reporter passed her driving test. And no one ever believes me when I say we worked on Computer Weekly with manual typewriters and carbon paper. Perhaps now they will. And fax machines the size of a small cupboard. Nice to see how the 'nursery' produced some great journos. I still freelance, but now mainly on testing and appraisal of caravans and motorcaravans and some industry consultations.

    Best wishes to you Susan

    [Personal details removed by Moderator]

  • Comment number 14.

    12. At 4:54pm on 31 Oct 2008, CobblyWorlds

    Thank you for this factual, objective information.

    I will view any other reports in light of this, but would still wish that every negative ecological story was not inevitably topped or tailed as being 'due to GW' (when my preferred, if less catchy, embodiment is 'Probably Man-Worsened Negative Climate Change') as this often leaves the topic at hand and its science open to diversionary criticism.

    This would seem not to have been the case here as the claim was to counter the food chain facts, but I still remain concerned that various entities within the BBC can issue rather diverse pronouncements (if one was news and the other a so far unsubstantiated claim by a listener read out by the DJ) within 15 minutes of each other.

    Many would be unlikely to ask, as I did (no answer from the programme makers to date) and maybe retain the last piece of 'information' more than the former.

  • Comment number 15.

    NOT SO FAST! (#10)

    "This energy imbalance is best explained by the Enhanced Greenhouse Effect caused by human CO2 emissions. i.e. Global Warming."

    What other explanations have failed your tests, against which you reference this finding? Is 'best' equal to 'correct'? Galactic electricity might play a part, have you had a look at that (Electric Universe)?

    Perhaps the sand eels are telepathic, read my mind, and just gave up.

  • Comment number 16.


    Get any alternate theory published in a peer reviewed (PPR) journal and I will consider - the alternate theories have now all collapsed and have no contemporary presence in PPR science. (Yes I have looked at Electric Universe - predominantly bunkum - my degree's in Electronics)

    As this is all off topic, I'm butting out.

  • Comment number 17.

    I'm obviously a bit late coming to this, but there is a solution to the problem of idle computers consuming power needlessly: go asynchronous!

    And there's no need to go all that way to California either. Here at the home of computing at Manchester University, our group has been working on chips that only consume power when they do something. And we've been so successful that many readers will already be using our chips in their mobiles.

  • Comment number 18.

    I'm sure there's a point here somewhere. As a former director in a company that designed and built data centers, this comes as no surprise to anyone in the industry. Did you think all that powerful computing happens by magic? We have lots of research which explains how to get the most efficiency out of our designs and operating practices to obtain the greatest bang for the energy buck. The US government, industry focus groups, individual companies and much more. Right now one major source of consumption is blade servers using multicore processors. A single cabinet can consume anywhere from 6 to 25 kilowatts. A six thousand square foot room can require over a million watts of power for the computers and another million watts of air conditioning. With redundancy in both commercial power and generator backup, it's an expensive proposition and an engineering challenge. For many users, these are mission critical facilities which cannot be allowed to fail. Therefore, even with all this redundancy, a backup "hot site" is often built just in case. A good source of information is The Uptime Institute. Nobody wants to be off line. No customers want to have service unavailable...ever. BTW, BBC's reliability in this regard has a dismal track record. Very bad.


BBC © 2014 The BBC is not responsible for the content of external sites. Read more.

This page is best viewed in an up-to-date web browser with style sheets (CSS) enabled. While you will be able to view the content of this page in your current browser, you will not be able to get the full visual experience. Please consider upgrading your browser software or enabling style sheets (CSS) if you are able to do so.