My carbon offsetting your server blog post has been getting a lot of search traffic.  I’d like to know peoples experiences of offsetting hardware resources so please take a minute to leave a comment.

My colo box will be going virtual pretty soon so I will oficially unplug it with any luck. My ipv6 transfer didn’t go well, so I’ll make sure the virtual box has IPv6 and migrate everything slowly rather than changing every DNS record and borking a few things (like my smtp filter) in interesting and exciting ways.

So 1234554321 has passed and so too has 1234567890, I raised a glass to both & hope you geeks did too.

Kudos to Chris Rowe for http://coolepochcountdown.com/ the music and fireworks did make me chuckle!

Aside from Mattress shopping and ipv6 migration for the rent-a-rack I’m sure St Valentine will suck-up plenty of my time today.

Have a great day.

How easy is it to parse this string ?

User-Agent: Mozilla/5.0 (X11; U; Linux i686; en…

Mac ?

Still looks like fun though (thanks to cnet), the point mapping functions could be be great for virtual viewing of houses Etc.:

Now turning the point map and images into vectors and textures could make it a near instant hit in the games geeks. Game of Quake in the V&A or the Louvre anyone?
GTA in the glowing areas of Amsterdam would be closer to the mark I think

I’ve been a twitter fan for a few months but with friends on such a variety of social sites it’s been a challenge to keep some form of status on the disparate services. Ping.fm has found and filled that niche by pinging a huge number of social networks and blogs too all at once, radio style.

It’s lacking an awesome client / firefox plugin. (I’ve been a twitterfox fan since day one). With an API already published I’m sure there is a twitterfox and twitterberry clone (google pingberry ) (spammers registered pingfox )

The biggest advantage however is the IM client. Twitters has been borked for months and only supported googles jabber thingie and nearly everyone I know are on the other three big networks. Ping.fm’s yahoo IM integration was the deal clincher as thats what I use on my Blackberry.

Use the ping.fm beta signup code “vivalaping” to try it out for yourself.

ReJaw is looking cool too, the way they include media in status updates I’m sure will be a very useful feature.

^{(1 paragraph version)}
Roughly speaking your typical geeks colo box uses roughly about 1Amp at 230volts – (1 cpu 1 disk 2Gb ram and a PSU thats a few years old) Thats about 2 tons of gold standard carbon offsetting you need to do to earn your self some Green Karma.

Geeks all have “their box” hosted at a colo at “mates rates” or a dedicated server hosted in a re-developed mall in central USA where power is dirt cheap. or hidden at the bottom of a rack in a building who’s name starts tele-something. They relentlessly use power and cooling 24×7x365.25 and have a pretty substantial impact on the environment as a result.

Mine is in Amsterdam, it’s a nice place to visit for a fun weekend

Generally it does enough work so justify its existence, however I feel the same about it being there as I did when it was at home in the loft. It’s sucking power and a bit of a geek luxury. So I figure I need some karma points and did some research into offsetting the CO₂ generated powering the thing. I don’t link M$ very often but this blog article explains PUE very well nice idea for a corporate blog too.

carbonfootprint.com say you generate 430g (grams) of CO₂ per KWh (Kilo Watt hour)

Traditional Power Station Emissions
Oil 740g/Kwh : Coal 32g/Kwh : Gas 385g/Kwh

430g/Kwh will do for me.

Your server has a reasonably lousy PSU if it’s over a year old, a hot CPU a few fans a squeaky disk or two spinning like mad and some RAM and a chipset – they draw about 1 amp, In the UK a 300W PSU = 1.3A and if your running at 100% load it won’t have lasted more than a couple of months so lets do the sums assuming your running at a single Amp.

1Amp = 230W
230Watts*24hours*365days (you hope) = 2014.8Kwh/year
2014.8Kwh * 430g/kwh = 866020g of CO₂ = 0.86602 Tonnes of CO₂

Thats only half the story…. you’ve got to cool the thing too if you’re in a colo.

Energy efficiency calculations are fought with pitfalls so here is an easy way to do it , it’s called PUE.
PUE (Power Usage Effectiveness ratio) = Total Power used by building / Power used by racked kit
(I told you it was simple)

Do you know the PUE for your data center? Nope? nor do I, however thanks to this UPS case study I’m stealing this factoid… “According to the Uptime Institute, the average ratio is 2.5”

We’re going to assume a PUE of 2.5, So 1 Watt used by the server is actually 2.5 watts of energy consumed by your colo’d box overall.

A 1 Amp server is responsible for 0.86602 Tonnes

Timesd by a PUE of 2.5 = 2.16505 tonnes of CO₂ Per year.

Wow

So don’t go out and plant a bunch of trees in your back garden yet because there is a better (for the planet) way to offset CO₂. Think about the Carbon cycle for a second… By investing in a project that saves energy in a place where they use only fossil fuels to generate power your offsetting monies have the greatest impact on the environment. It works like this. Your 1 amp little server uses the same power as about 10 60W incandescent bulbs. CF bulbs are 80% more efficient than incandescent bulbs so you need to replace 12. Replacing incandescent bulbs in areas that use oil is maximizing your offset investment for the planet to the tune of 740g per Kilowatt Hour for oil fired power stations.

If you’d rather offset locally than thats still a good thing to do

So round it up to 3 Tonnes if you need the Karma to offset the extra 5” of Flatscreen space you don’t use because your hacking in a 80×25 ssh terminal all the time

co2balance.uk.com have an interesting set of projects with good descriptions too.

Justin posted about some worthy offset projects too.

Please make sure that your offsetting is done via a non-profit organisation and that they meet with both the Gold Standard and Voluntary Carbon Standard to maximise your good deed.

Let me know if you did below!

Pete Moore originally recorded “Asteroid”, the Pearl and Deane theme in the 70’s and it’s been my ringtone for about 5 years. It’s short and lively and no other bugger has it

Having the Blackberry ring like a classic 2 bell phone was just not right so I went on the rampage to find Asteroid once again. The Pearl & Dean site has links to iTunes but thats not very helpful on the Blackberry or Linux but the flash app on the right hand side of the page loads mp3. Superb! So here are the links.

Asteroid Now! Asteroid 2006 Asteroid 70’s * 60’s music 50’s

[right click & select save as]

I can’t decide on the new one or stick to the 70’s one though.

O2 are no longer stocking the Blackberry 8320 and now only stock the 8310. The difference? The former saves your bill by having wifi. Why don’t O2 want you using wifi? Oh gee I wonder.

Living in a village with municipal wifi won’t be saving my 3G traffic bill if we upgrade directly with O2 then. (And their internet package AUP forbids voip too.)

Cheap trick O2. Nice way to treat your customers; Down right scummy. 

One of the most popular things about firefox 3 is that the plugins provide so much useful extended functionality. This has got to be heavily responsible for it’s phenomenal success.

I was reading firefox_answers on twitter earlier and it’s like a broken record. Plugins are to blame for everything apparently, yet there is no way of knowing which cause you resource issues when FF leaks like the titanic. It’s the fastest browser because X & Y say so Yawn (Excuse me).

Well I’ve not taken the pill yet, I’m running the portable version currently to test the waters. But one thing I’ve not noticed is this apparent speed demon of a browser now it’s set up the way I like it. So, I’d love to see the test results in a common configuration rather than minimal configuration with a handful of the most common plugins installed. Adblock with the juicy lists enabled, a search toolbar, download status bar, realplayer, bookmark manger like del.icio.us and flash enabled. Thoughts?

I’ve got a box that runs a bunch of smoke pings 24×7 and I have to admit I look at it about once a week. However guilt got the better of me and I decided it was about time to upgrade Ubuntu from a recently up to date v6.06 LTS to v8.04 LTS and I took the JFDI method “sudo update-manager -d” after copying off the smokeping config off box. Well what a surprise this was this morning:

The Standard deviation improvements are right after the upgrade were a pleasant surprise, especially considering this isn’t a local box I’m pinging too.

Got any idea why ? Leave a comment…

Having spent an interesting evening chatting with Dave from MailChannels, I’ve now spent well over a week traffic shaping SMTP for some 200k connections and it’s time for a few observations:

1. The installer is very Solaris-a-like. I’m not a fan. I do remember a few far worse, but only a few


2. Upgrades are somewhat clunky, since the installer A) takes settings saved in ~/.something rather than the running config (weird) & B) buggers about with your config file removing all those logical line breaks you put in to define sections once you enabled some of the disabled features.


3. I did wonder why the hell it wasn’t a direct replacement smtpd for postfix at one point. XCLIENT functionality might cut the mustard though.


4. FP’d disastrously on a finance site I use quite a bit. I had to whitelist it


5. It annoyed my kids by FP’ing on a very popular music site they use a lot


6. >90% of rejections were RBL related.


7. I’ve a suggestion list a few pages long, but raised enough bugs already. I’m not really a grumpy-old-man


8. It appears not to like the popular catalog shops, this isn’t so much of a problem because I can’t afford the wifes shopping habits at the moment.


9. There is a setting to turn off the feedback & Ken made a change to the license to exclude addressing info but my preference is to opt out on my personal box


10. It appears to be catching less (or I’ve been noticing a lot more) and I’m relying on SpamAssassin more since the upgrade

I’ve a hunch that most of these issues come down to bulkers postmasters setting short timeouts and increasing parallelism to get the greatest bang-for-the-buck-come-instant-gratification from their listservs in the zero hour. Postfix and qmail come with excellent defaults that cope with sods like me greylisitng or traffic shaping spammers connections – don’t fuck with them eh?

This is after all my personal mailhost, it runs a few small spamtraps and a bunch of mailboxes but not scientific empirical datasets. My mailbox has been hard to manage this week and I’m not aware of anything out of the ordinary happening. My public self seeding site is getting more english junk: This week then the same week last month. Pure observation but interesting none the less.

If I had to sum it up: At the moment, I’m sorry to say it’s not as effective as the decent greylisting implementation I was running, it appears to be FP’ing more but nevertheless it has far fewer drawbacks in normal use since the delays appear more usual than the arbitrary 3^rd party retry timeouts greylisting causes. It could be excellent. (Edit: added the dot. Is anyone else trying it? )

The Granddaddy of rejections was very interesting though natwest.co(m|.uk) WTF is going on with all that phish?