On Fri, May 13, 2022 at 05:38:13PM +0200, Thomas Goirand wrote: > On 4/8/22 20:35, Julian Andres Klode wrote: > > Sponsors: When receiving sponsored resources like electricity, we > > should inquire about the carbon footprint of those resources, and > > what the sponsor's approach to environmental affairs is. > > Most of the time, "green energy" is just "green washing". If you buy "green > energy" in France or Swiss (these are the only places I know for sure what's > going on), you get a higher electricity bill, and a slot in the green energy > consumers. But the electricity may well come from the nearby coal power > plant, even if you bought a slot of green electricity. Proof? References? I'm in Germany, and for the time I got to decide, I chose a 100% renewable electricity provider. I'm happy to know that my bill goes to someone deploying renewable generation instead of coal plants. I do believe them. If they get caught cheating, they can say good-bye to their business. Whether "the electricity comes from the nearby coal plant" -- the electrons never leave the building, as you might know [0] :-) It's about what investments your bill goes to, as I said above. > IMO, you're much better off fighting this at another level: lobbying your > government to do what you think is right. We gotta do both, I think. > > Budget: We need to determine our current CO2 emissions as a project, > > and then define a roadmap to carbon neutrality by an acceptable date, > > I think 2035 or 2040 are commonly referenced. This is likely to be > > exponential. We should use project funds to hire an expert consulting > > firm to do this for us. > > If I had my say, I would vote against (wasting) money for such an expert, > and wasting contributor time on this. I'm tired of reading about CO2 > emission in the data center, when old servers are just trashed, and when > electricity production is out of the control of the data center owner (see > above). You just put some assertions without backing them up in any way. And, by the way: perhaps there are experts out there willing to do some pro bono work for Debian. > I pushed my company to recycle old server and re-use them as long as > possible, and we went from a 10 years lifetime to 15. That's IMO a much > nicer and efficient approach for protecting the environment than just the > green-washing CO2 propaganda. Another unwarranted slur. A computer has a lifecycle, and it makes sense to take all into account. It does make sense to re-use for some time, but it also does make sense to take the power consumption into account (more modern servers tend to have better MIPS/power ratios). This [1] source (alas, not dated -- copyright is 2022), for example says that energy use is 34% of a computer's lifecycle. > > Monitoring: Once we have determined our CO2 emissions and defined a > > roadmap, we need to constantly monitor our CO2 emissions to make sure > > we stay on target. I propose quarterly environmental impact reports. > > A quarterly environmental impact that only takes CO2 into account is only > part of the reality. Do you have any idea about the environmental impact of > mining these rare minerals needed to produce a server? Another example: > producing the aluminum needed for a server chassis use a huge amount of > electricity. Well, "huge" is not a number. There are studies out there (I tried to provide an example) which try to get down to more concrete figures. I think actually trying to assess "where are we" and "where do we want to go" is a very commendable goal. I'm happy that people are trying. Cheers [0] If I remember correctly, electrons in copper move in the rough ballpark of 10^-6 m/s, so at 50 Hz they just wiggle around about 1/50th of a micron. But my physics are pretty old, so glad to be corrected :-) [1] https://sustainablecomputing.umich.edu/knowledge/life-cycle.php -- t
Attachment:
signature.asc
Description: PGP signature