Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Well, the marginal environmental cost of me doing another Google search is pretty close to zero -- those servers are already running anyway.

How many requests/sec per server do you think Google handles? Then we can try to guess how many additional queries we would need to make for them to buy an additional server (causing higher power consumption).



>those servers are already running anyway.

The power plants are on anyway, I might as well leave my heat on and windows open.


Not the same thing.

The energy consumed by your home scales (roughly) linearly with usage. The energy consumed by a computer does not scale linearly with the number of http requests -- most of the energy is just to keep the machine up & running, with a minimal additional cost per request.


But more searches requires more servers in the first place.


Yes, but it's not linear. It looks more like a "step function." (http://en.wikipedia.org/wiki/Step_function)

If you currently use a heater for 1 hour/day at home, and then you turn it up to 2 hours/day, you do need about twice as much energy to power that heater as before.

However, if I make 100 additional requests today, Google doesn't need a new server. If everyone in the world makes 1% more requests, Google doesn't need a new server. So to try to calculate the cost of me doing one additional search is perhaps slightly misleading.

I'm not saying there is no cost. I'm just saying that the cost is incurred in chunks, and not "per request".




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: