How big an eco-hazard is IT equipment?
A hard charge to make stick...
Are computers bad for the planet? As a consumer of power, IT equipment is certainly all too visible, and is shaping up to become a prime whipping boy for governments striving to get a lid on their CO2 emissions. After anything that burns petrol, PCs, chargers and that arch eco-criminal, standby mode, now come pretty close to the centre of the crosshairs.
And they're being targeted by Brussels. This year, as part of its action plan for energy efficiency, the European Commission will begin to enforce minimum energy performance requirements on a range of electrical appliance categories, including IT equipment and consumer electronics. It has selected 14 priority groups of products which will be subject to energy efficiency labelling and the somewhat more complex EcoDesign Directive.
The labelling end of the deal is relatively straightforward, giving an indication of how much juice an appliance uses to do the job it's supposed to do, whereas EcoDesign is intended to set down broader design requirements for products, dealing with "the consequences of energy consumption, consumption of other materials/resources, waste generation and release of hazardous substances to the environment" and involving "systematic integration of environmental aspects at a very early stage in the product design". (more details here).
At its best, an EcoDesign product could be said to be "best of breed", environmentally speaking, while at the very least EcoDesign will define maximum consumption and efficiency limits for products sold within the EU. By attempting to specify what is and is not acceptable in product manufacture and design, however, it could clearly be controversial, and spark international rows over trade and protectionism.
Energy efficiency labelling is more readily achievable. Among the 14 priority groups nominated are computers, TVs, chargers and PSUs and "standby and off-mode losses", with "special attention" being paid to standby losses.
These are a popular target, and were also of particular interest to MEPs recently, who in addition to demanding the abolition of patio heaters, called for the commission to set a one watt standby mode limit, and to research the likely benefits of moving to zero watt standby.
No sooner said than done. When it comes to environmentally aware computing, Fujitsu-Siemens Computers (FSC) is one of Europe's old-stagers, and last autumn announced zero watt standby monitors, which use capacitors to store sufficient charge to kick them back into action.
Speaking to The Register earlier this year, FSC CTO Dr Joseph Reger unpicked the economics of standby mode. A monitor with a regular power adapter would consume approximately €1.60 worth of power per year, so switching over to zero watt standby, assuming 16 hours of standby time per day, would save less than €1 per monitor per year.
Which, as Dr Reger readily concedes, is not exactly a compelling argument for the immediate replacement of existing kit with low or zero standby mode devices. Financially the "savings" could easily be eclipsed by relatively minor discounts, and manufacturing and recycling costs would be greater than the power consumption-related benefits.
But, as Dr Reger points out, if you could have a zero watt standby monitor for the same price as a 1 watt one, why wouldn't you go for the zero variety?
The key in this part of the field therefore lies in producing low power, more environmentally friendly products at standard price so that they become mainstream, and - here he's clearly in agreement with the commission - making people more aware of their energy footprint via energy labelling. People will go for the lower power product so long as it's not more expensive, and so long as they actually know it's lower power, hence the importance of marking them.
It's a similar story with chargers and PSUs. Power wasted by a typical notebook computer or monitor is approximately 12 Wh per day, which Dr Reger estimates would amount to 12.74 kWh per year, worth approximately €1.27. A similar saving could be made by simply using one 100 watt light bulb for 30 minutes less every day. So implementing a company policy of turning off the lights after everybody's gone home, or better still whenever an office isn't being used, would be vastly more effective than getting worked up about whether or not a charger or PSU is plugged in.
The efficiency of external power units is an issue in addition to "waste" or consumption, and here there are gains that can be made in percentage terms, although again these won't necessarily amount to much in terms of cold, hard cash.
The bulk of external PSUs currently on the market are between 80 and 85 per cent efficient, FSC's current benchmark being 85 per cent. Going much above 85 per cent at the moment would be significantly more expensive (87 per cent is achievable at a premium), and while most manufacturers' equipment exceeds 80 per cent today, not all does.
Efficiency ratings are complicated by the question of load, with PSUs being at their most efficient when under full load. They will also have a "no-load" power consumption level when plugged in, this being 0.5 watts (i.e., negligible) in FSC's case. Load calculation doesn't have a great deal of relevance to monitors, which are broadly on, on standby or off (although brightness setting will have some effect), while the onboard battery of a notebook computer means the CPU load is independent of the PSU load, so your calculations get more complicated.
Although these numbers don't add up to an argument against power efficiency, they put the commission's targets into perspective. It certainly makes sense to develop standard price technology that's more power efficient, but the greatest effect of the commission's energy efficiency plans will lie in raising public awareness of their consumption - the savings to be had from the carbon footprint of individual items of IT equipment are relatively small.
Grasping this might also be beneficial to those elected representatives who're obsessed with "waste" caused by low-impact devices almost to the exclusion of bigger ticket items elsewhere.
"The waste from good external PSUs for notebooks and monitors is almost negligible," says Dr Reger. "Further improvements there generate only marginal savings. But there are still a lot of bad PSUs around in the field or offered by vendors who put cost over responsibility that need replacement. So the public hype around PSUs is justified for old stuff, not for good current and future external PSUs."
Commission figures (for 2004) put this into a broader perspective. In the pre-enlargement EU states, six per cent of power consumption in the home could be ascribed to "consumer electronics and other equipment standby", three per cent to TV and one per cent to office equipment. Heating (including water and aircon) was 36 per cent, lighting 12 per cent and white goods 31 per cent. The power monsters are all to obvious.
The relatively high share white goods have of the total take is particularly significant in understanding the objectives and likely impact on the commission's plans for IT equipment and consumer electronics, because these are intended as a follow-up to earlier drives for energy labelling of white goods. Washing machines, tumble dryers and fridges clearly do use large amounts of juice, so a two-pronged attack on them - flagging consumption levels to consumers and pressuring manufacturers to improve performance - could, and did, achieve significant gains.
Any legislator expecting major direct gains from little black boxes that sit in the corner feeling vaguely warm, however, is going to be severely disappointed. Faced with the information that a plugged in Nokia charger is costing them a fraction of their annual cappuccino budget, consumers are unlikely to be particularly troubled.
More usefully though, faced with the information that their lightbulbs eat a lot more of the planet than their chargers, they'll be more likely to focus on the things that make a more serious difference. IT and consumer electronics could end up as the poster child of environmental responsibility, as the numbers start to become more public.
With a couple of exceptions. The switch to digital has, in Dr Reger's view, has had a significant effect on overall consumption, and of the appliances on the commission's hit list, the plasma screen is very much the Hummer in the living room.
An old style CRT TV (older readers might remember these) might use about 100 watts, while a sizeable LCD unit could use double that. Plasma displays, however, can tip the scales at anything up to 500-600 watts, so a biggish home cinema system could be costing something in the region of €200 a year.
It's consumption that barely existed ten years ago, and now it's a significant percentage of total domestic power consumption. Nor, comparatively speaking, is standby the big issue here - it's what the things do when they're on that's the problem.
The data centre is an exception that isn't directly addressed by the commission's IT plans, which focus mainly on specific devices rather than collections of them with associated environmental control.
According to Intellect (High Tech, Low Carbon), data centres use 2.2 to 3.3 per cent of the UK's total electricity, with 50 per cent of this being accounted for by cooling systems.
Big iron doesn't do standby terribly well, and even if it's just sitting there doing nothing and waiting for something to happen, a data centre is still going to be racking up a fair old tab. If at zero load you're using 50 per cent of the power you use at 100 per cent load you're doing good, and if you find yourself running at 100 per cent load all of the time (which is where you might deem the operation most 'efficient') then the career-sensitive IT manager is going to get some more spare capacity in there, fast.
At the individual hardware level switching over to greener processors that use less power and generate less heat makes some difference, while smart cooling systems can reduce the level of spend on climate control. Beyond this, consumption reductions can be achieved via the potentially contradictory routes of consolidation and decentralisation.
By viewing server capacity as a centralised resource and employing virtualisation, it's feasible to reduce total server capacity by making 'on-demand' capacity available. Alternatively, smaller, low-noise, low-energy servers that don't require a separate server room could be used by SOHO customers to reduce or eliminate the need for centralised server capacity.
FSC offers both routes - lightweight SOHO servers and units with VMware built-in, while Dr Reger sees virtualisation as a route that leads to substantial savings.
Given the complexity of the issues involved, however, IT managers are more likely to get beaten up about overall IT power bills than they are to get a clear understanding from bosses and legislators about what makes sense and makes a difference, and what doesn't. Big, quick, wins are therefore necessary, and the data centre looks a fruitful place to achieve them.
Consider, though, what it is that makes IT a power villain in the first place. Why, as Dr Reger puts it, is it fair that IT should start from zero? The technology we've deployed over the past 30 years has to a great extent been intended to produce savings, and if we think in terms of carbon footprints, then a pretty substantial IT green revolution could be said to have taken place before it even became fashionable. A little halo-polishing is therefore, possibly, permissible. ®