Four aspects of Green Computing December 11, 2006Posted by ficial in brain dump, environment, green.
Williams has recently been at least vaguely interested in Green Computing. There were a number of articles a while back about it in the Chronicle of Higher Education, and it was mentioned at a few conferences I and/or my co-workers attended, and it’s a topic in which I’m interested so I push the idea a bit around the office and larger campus. Finally, the college has recently begun to more closely monitor energy consumption and various people were astonished and a bit panicky at the amount of energy the computer center is using. Computers are un-green in three to four (depending on how you count and group them) main ways: the energy they use (this generally gets the most attention), waste when they break or are superceded (this usually gets the second most attention), the habits they engender (usually ignored), and their embodied energy (also usually ignored, though sometimes considered as a part of the e-waste discussion). Those problem areas are a good place to start when thinking about making computing more green.
Energy consumption is relatively straight forward. You can get a simple electricity meter (e.g. Kill A Watt) and figure out how much energy a computer system is using in various states (what the computer is doing – word processing, listening to music, running a ‘screen saver’, etc.) and configurations (what kinds of parts make up the system – what size screen, speakers, printer, etc.). Here are some rules of thumb about using as little energy as possible when computing:
- the fewer peripherals a system has the less energy it requires
- more powerful machines require more energy (sometime much more – e.g. the Mac G5 workstation uses 185 watts when idle, while the iMac G4 uses 46)
- larger screens use more energy than smaller screens
- flat-panels use less energy than CRT monitors (though not so much less that it makes sense to actively junk CRTs and replace them with flat panels)
- laptops use much less energy than desktops (generally they use less than a third the energy of a desktop, but they’re more expensive, more fragile, and usually have a shorter useful lifespan)
- the harder a machine works the more energy it uses (playing a video takes more energy than editing a text file)
- the more machines you have in agiven space the more your secondary energy drain will be for keeping the space climate conditioned (though if your building is smart enough to have a room-by-room HVAC system you may be able to recoup some of that energy in the months for which you need heating).
- energy star rating is good
- more functions per piece of equipment is good (from a single monitor attached to multiple computers, or 18 virtual machines running on a single piece of underlying hardware)
- machines should always be running at the lowest acceptable power state (sometimes a machine needs to be fully on 24/7, but other times it could get away with only being on a couple hours a week and should be off the rest of the time)
It’s worth noting that most of these are trade-offs: a system with out a printer and scanner will use less energy, but some people need printers and scanners for their work; a 10″ screen (which might minimize energy use) for a standard desktop is just as ridiculous as a 28″ screen (though for different reasons), etc. Green computing isn’t about absolutely minimizing the energy used by a computer system (for that, just leave it unplugged, or simply don’t buy it to begin with), but about finding the right balance. This machine might need a scanner, but perhaps it doesn’t really need speakers, a web-cam, nor a floppy drive, and maybe a 17″ monitor could be used instead of a 21″. This aspect of green computing is usually a pretty easy sell because it (usually) saves money both up front and over time.
Electronic waste is a much more complicated issue, and I know a lot less about it. The high points as I understand them are:
- electronic waste is often shipped to other countries where it is buried unsafely or burned
- a number of manufacturers are now running recycling programs, and you should participate if at all possible
- the faster new computers are bought the faster the old ones turn into junk
- donating old computers to schools or non-profits doesn’t solve the problem, it just pushes the problem on to the recipient (that doesn’t mean that old machines shouldn’t be donated, just that donating them doesn’t mean the eventual disposal problem can be ignored)
There’s no easy solution to electronic waste. (At least, none within the current manufacturing paradigm. Concepts wil crable-to-cradle design and build-to-rebuild would go a long way towards remedying this problem, but for now they’re not common enough to count on, especially in the electronic industry). I think the best option these days is to send ‘used up’ machines to a recycling program, which costs money and so can be a hard sell.
Computing habits are nice in some ways because they require no technical changes / advances, but they’re notoriously hard to change. Some of them arise from things that were true in the very early days of computing and don’t apply with current technology (e.g. ‘screen savers’, leaving a computer on so it doesn’t go through a startup cycle), and others are simply behaviors that people have developed for whatever reason (e.g. printing out documents instead of reading them on-line). People in all sorts of realms have thoguht long and hard about how to change people’s habits: advertising firms, governments, schools, activist groups, religions, etc. It’s well beyond the scope of this post to go over that in detail, but in broad strokes, it’s about marketing/convincing, educating, and restricting/forcing. In green computing this means telling people what to do in posters, in person, on screen, etc., explaining to them the effects of what they’re doing and why it’s important, and providing machines with good default settings (eg. screen turns off after 10 minutes of inactivity, always print duplex, etc.) and maybe locking them there in some cases. Some problem behaviors of note here are:
- leaving computers on 24/7 – most people really don’t need it, and modern computers aren’t harmed by it at all.
- leaving monitors on all the time – monitors use quite a bit of energy and have a very short startup time. Even if someone is stepping out of the office for 15 minutes it’s probably worth getting them to power off their monitor. Also, modern monitors have good power management options, take advantage of them.
- running ‘screen savers’ – a screen saver does nothing but eat energy, and for many people it’s the most energy demanding application they run. To really save the screen, have it go to sleep or turn it off.
- peripheral explosion – not every computer needs all peripherals. Only include the ones that are needed, and for those make sure they’re off when not in use.
- printing lots of everything – we have serious printing problems, with people printing multiple copies of large/long documents which they never even pick up from the printer. In general, avoid printing if possible and if printing is necessary it should be duplex (both sides) if that’s reasonable and should only be of the needed pages.
- leaving machines/peripherals on when no needed at all – this is a bit different from the first point. The problem here is more with people not thinking than with people thinking the wrong thing. If someone’s going on vacation for a week, they really should turn off their system before they go, and maybe unplug it too (to avoid a vampire load)
There are many other such things – I’m sure everyone can think of a few. Some occur at the level of an individual (e.g. a person who habitually prints out every email), some at the level of an organization (e.g. computer labs that are left fully powered when the students are on break), and some societal (e.g. users have been trained that any computer they sit at will be on, and if it’s not they’ll call technical support before trying the power button). There’s a lot of work to be done in this area, and none of it’s easy.
And finally we get to embodied energy. If you don’t want to read the detail in the wikipedia entry, you can think of it as the total energy required to create a thing, from initial resource extraction and refinement, to transport, packaging, etc. The whole shebang. There are other approaches to the same idea (emergy, system ecology, footprint, etc.), and the particular one used doesn’t matter that much for my purposes. They all get at the idea that the dollar cost of a thing is not necessarily a good measure of the impact that thing has on the planet (though many of the methodologies mentioned are being used be economists to try to internalize some of those external costs). Embodied energy is why, for example, it doesn’t make sense to go out and replace all your CRTs with LCDs. The flat panel does indeed use less energy, but over the expected life of the equipement you will not save enough energy to make up for the energy it took to create that CRT you just threw away. Unfortunately there’s little in the way of solid information about the embodied energy costs of things. This leaves little to go on besides generalities and intuition. For now, that will have to be enough.