Recently in environment Category

Over the past month or so I've been quietly posting on a new site of mine, iveswitched.co.uk. It is a microcampaign site encouraging people to replace their power strips

The power supplies that so many of us use to support our digital lives all drain power when they are plugged in, even if their device is not actually switched on. You can feel the warmth of them if you put your hand on them. Much more detail on iveswitched.co.uk, but the take away message is to assess the things you plug in and consider turning them off at the socket. A power strip with switches can help you do this more easily.

After seeing Gavin Starks speak about AMEE at Xtech I've been thinking more seriously about the amount of energy that everything I do uses. I'm reminded of James Duncan Davidson and Chris Messina, both of whom have written about green issues.
I think that as web developers we should think carefully about the amount of cpu cycles and bandwidth we use in our applications. We should be optimising both our usage of energy and the performance of our applications. I'm not advocating a hairshirt attitude, we still need to serve up pages to our users, but we can be a bit more thoughtful in how we do this. Yahoo take this seriously, too.

At the @media2007conference I asked the panel at the end about the environmental impact of what we do and got somewhat blank looks from most of the panel. So I think we need to do a bit more discssion about this amongst tech circles. Look at what you send to and from the client, AJAX can help a lot in terms of avoiding full page refreshes, but watch out for accessibility issues. Use mod_gzip on your web servers, ensure that you support the not modified 304 http status on your feeds. Look at asynchronous means of responding to queries, Jabber will have an interesting role to play in this. Many bits of information do not need to be processed live and immediately returned to the client, some do I'll admit.

A lot of the tips for improving performance will help save the planet too, so at least there is a primary driver for companies to implement this. Something I'd be really interested in would be the trade off between cpu cycles and bandwidth in terms of energy usage. Is it better to gzip and then decompress to avoid sending excess data through many routers, or does the gzip and decompress use more energy? It seems to make sense that sending less data is more efficient.

Textdrive have been exploring this area, as part of their move to Solaris and Sparc machines is to lower their electricity bill. So, I'm pleased that I'll be moving to one of their new solaris machines soon. Amazon EC2 can also be seen as a good initiative in this area. Why have your own capacity sitting idle just in case. Much better to have your code swapped in and out on another machine which is being otherwise kept busy.

I know that this may seem a tiny issue, compared to leaving power adaptors plugged in and electronics in standby, but every little counts. If you have a site then every excess cpu cycle or byte you send uses power every time you serve up a page. Whilst each page downloaded is a tiny contribution, there are many of them. So keep this in mind as you make new websites and applications.

Tags:
enviroment, web development , climate change, performance, bandwidth

Yesterday I was thinking about the disclaimers appended to emails, oddly the one from @Media, the 832 bytes of disclaimer didn't really say anything helpful. They do not have a strong legal position which there questions their enforceability. What bothers me is not that they include the disclaimer, but the amount of data that needs to be sent, stored and probably backed up by many people.

I guess some lawyer said that it was more enforceable if it formed part of the same document and was not a link to a standard disclaimer. The plaintiff being unable to plead to not having read the disclaimer. A UK perspective, which is also generally not in favour of them.

So lets do some maths, call the average disclaimer 512 bytes to make the maths easier and underestimate the final figure. The BBC and Macmillan both append these automatically to outgoing email, as they are my former and current employers lets go with them.

512 per email times say 10,000 employees sending 200 external emails a month equals 1024,000,000 bytes of additional date or nearly 976MB of extra data sent and stored around the world. Call that 12G a year, not a lot you might think, but that is a tentative estimate for the BBC alone and we are at a gig a month already.

Data requires electricity to move about, to be stored, backups use more electricity and then energy to be moved off site. How much energy is being used in shifting just these disclaimers around the planet?

I'm certainly going to be asking my employer to review the use of them, I suggest you do the same.