Solar powered sites are cool and fun, but I find it ultimately lacking because so much of the rest of the networking infrastructure is reliant on the grid. It would be more energy efficient to just host the static site on cloudflare or whatever, and use the solar panel to charge some batteries, or something you would normally use the grid for. I suspect overall energy usage would be even lower if the site was hosted on a CDN, due to the CDN operators keeping their machines near full utilization, and fewer network hops required for an average request.
He's looking at it from a utilitarian point of view.
From that perspective, you've got a finite amount of energy that's being consumed at any point of time and by that logic, you'd be better served to use the solar energy for something else in your household that would've needed the grid, because a cloud provider will almost certainly only need a fraction of the electricity you'd need to keep the static files served.
So, by doing that you've effectively reduced your absolute energy footprint.
But these projects aren't about reducing your energy footprint. They're about having a local webserver that's running on solar energy. Or in the authors words: At its heart, this project is about learning and contributing to a conversation on a greener, local-first future for the web
Yeah, it’s cool to just see what goes into it, and get a more intuitive understanding of the tradeoffs and what needs to be done to move more of this to green energy. It makes the (large) challenge of matching generation to consumption a whole lot more real. It’s obviously not meant to be a maximized carbon reduction in any way commensurate to the effort.
Agreed. When I see this type of thing, I am always turned off by people describing it as "greener" or “more sustainable”. Every small website having its own solar panel and hardware is not greener. People frequently think only of the carbon emissions of the energy used by the hardware once it's running, ignoring the carbon (and raw material) cost of building that hardware.
Serving websites is an area where capitalism’s promise of achieving efficiency of resource utilization through economic incentives probably actually works, via shared hardware.
This is a hobby and aesthetic thing, which is valid and interesting.
If anyone has some good data about carbon emissions of self-hosted vs shared hardware I’d love to see it.
> People frequently think only of the carbon emissions of the energy used by the hardware once it's running, ignoring the carbon (and raw material) cost of building that hardware.
The carbon is hard to account for in manufacturing. Solar, for example, is pretty close to being produced entirely with electric consumption and very little required CO2 output (except perhaps in the transport of silicon and other raw materials). The big energy draw for solar and battery is a kiln stage in both. Solar has to melt down the silicon which requires a high temperature furnace and batteries are basically "cooking" the raw materials onto their foil.
The math for solar is something like 1 to 4 years of generation before it pays back the manufacturing power debt. Batteries tend to be much shorter as they take less energy in their manufacturing process (with some hopeful techniques in the future significantly reducing that number).
Now, none of this is to contradict you, just putting the numbers out there. I completely agree that a server farm is likely to be far more efficient for hosting a website than home built solar powered pi. The CO2 emissions will be hard to beat, particularly if your cloud host resides in the PNW where power is nearly entirely renewable already.
There was a project a while back, sadly it seems to have gone offline, it was a kosher news aggregator. They had the most interesting rules, like only charging the batteries on Tuesday. It was such a weird and interesting projects.
Pretty cool. I use one of these small panels as a battery tender for a generator. It only is about 30w, but keeps a small 12v battery from dying over the months. It also has a charge controller built in - https://www.amazon.com/OYMSAE-Portable-maintainer-Cigarette-...
I suggest a different reasoning: what's about domestic p.v. with storage and racks in the basement with "free" A/C in terms of WFH and distributed "datacenters"?
How many have realized how much stuff can be hosted at home with availability levels not really far from most common datacenters?
Now, Low Tech Magazine also has instructions to convert a stationary exercise bike into a human powered generator, which you could build to add power during the winter :-)
> When I started this solar-powered website project, I wasn't trying to revolutionize sustainable computing or drastically cut my electricity bill. I was driven by curiosity, a desire to have fun, and a hope that my journey might inspire others to explore local-first or solar-powered hosting.
> The cost savings? Looking at our last electricity bill, we pay an average of $0.325 per kWh in Boston. This means the savings amount to $2.85 USD per year (8.76 kWh * $0.325/kWh = $2.85). Not exactly something to write home about.
I can't speak for his specific apartment, but apartments in general are more energy efficient that a suburban house. New York City is one of the lowest per capita CO2 emissions in the US.
These figures are generally based on consumption rather than production for that reason. It’s not helpful to credit emissions to the people who largely aren’t driving the output.
City living brings efficiencies from operating at scale - fewer miles by car, block of apartments instead of single family homes.
How does it help the discourse to bring this up? Would you like it if someone inserted this kind of (irrelevant) and unsubstantiated information when commenting on an article you wrote?
Solar powered sites are cool and fun, but I find it ultimately lacking because so much of the rest of the networking infrastructure is reliant on the grid. It would be more energy efficient to just host the static site on cloudflare or whatever, and use the solar panel to charge some batteries, or something you would normally use the grid for. I suspect overall energy usage would be even lower if the site was hosted on a CDN, due to the CDN operators keeping their machines near full utilization, and fewer network hops required for an average request.
You don't like that the rest of the infrastructure is on the grid, so you're proposing moving the site to the grid too?
He's looking at it from a utilitarian point of view.
From that perspective, you've got a finite amount of energy that's being consumed at any point of time and by that logic, you'd be better served to use the solar energy for something else in your household that would've needed the grid, because a cloud provider will almost certainly only need a fraction of the electricity you'd need to keep the static files served.
So, by doing that you've effectively reduced your absolute energy footprint.
But these projects aren't about reducing your energy footprint. They're about having a local webserver that's running on solar energy. Or in the authors words: At its heart, this project is about learning and contributing to a conversation on a greener, local-first future for the web
Yeah, it’s cool to just see what goes into it, and get a more intuitive understanding of the tradeoffs and what needs to be done to move more of this to green energy. It makes the (large) challenge of matching generation to consumption a whole lot more real. It’s obviously not meant to be a maximized carbon reduction in any way commensurate to the effort.
the grid is going to continue to get efficiency and renewable energy improvements that help at a much more massive scale.
it's 1000x more helpful to just like only run your big appliances when your grid is using more renewables.
Agreed. When I see this type of thing, I am always turned off by people describing it as "greener" or “more sustainable”. Every small website having its own solar panel and hardware is not greener. People frequently think only of the carbon emissions of the energy used by the hardware once it's running, ignoring the carbon (and raw material) cost of building that hardware.
Serving websites is an area where capitalism’s promise of achieving efficiency of resource utilization through economic incentives probably actually works, via shared hardware.
This is a hobby and aesthetic thing, which is valid and interesting.
If anyone has some good data about carbon emissions of self-hosted vs shared hardware I’d love to see it.
> People frequently think only of the carbon emissions of the energy used by the hardware once it's running, ignoring the carbon (and raw material) cost of building that hardware.
The carbon is hard to account for in manufacturing. Solar, for example, is pretty close to being produced entirely with electric consumption and very little required CO2 output (except perhaps in the transport of silicon and other raw materials). The big energy draw for solar and battery is a kiln stage in both. Solar has to melt down the silicon which requires a high temperature furnace and batteries are basically "cooking" the raw materials onto their foil.
The math for solar is something like 1 to 4 years of generation before it pays back the manufacturing power debt. Batteries tend to be much shorter as they take less energy in their manufacturing process (with some hopeful techniques in the future significantly reducing that number).
Now, none of this is to contradict you, just putting the numbers out there. I completely agree that a server farm is likely to be far more efficient for hosting a website than home built solar powered pi. The CO2 emissions will be hard to beat, particularly if your cloud host resides in the PNW where power is nearly entirely renewable already.
except all of this could be used hardware and panels. no need to buy new.
That doesn't seem to make sense.
Yeah, it's only a 50W solar panel. The solar power could be upgraded to a 400W no problem and the issue you have will no longer exist.
Often people smart with a small system to get things figured out how they want it before adding more capacity. Seems reasonable, no?
Networking equipment shouldn't be much power, speaking from personal experience.
It can be on a small UPS, even one from Costco to run it all off battery, and then solar to refill.
Heavily inspired - I suppose - by the oldest (2018) solar-powered website, Low Tech Magazine https://solar.lowtechmagazine.com/about/the-solar-website
The solar-run site is at https://solar.dri.es/ in case everyone wants to check it out all at once ;)
Of course this got posted on a sunny, cool and breezy morning in Boston. Sneaky!
Is this the same page that had the background effect indicating level of the battery?
My guess is that you're thinking of https://solar.lowtechmagazine.com/
How many more panels/bigger battery would your need to move the switch and other required home infra to run the site?
Shame the solar panels can't do "self-healing".
I this this is cool, I know a couple folks who got homelabs on reddit who mainly use solar power due to the cost and want to go green.
This is so cool! I'm so surprised I've never heard of someone doing this before :)
There was a project a while back, sadly it seems to have gone offline, it was a kosher news aggregator. They had the most interesting rules, like only charging the batteries on Tuesday. It was such a weird and interesting projects.
Archive of the website: http://web.archive.org/web/20200707101320/https://jewjewjew....
Pretty cool. I use one of these small panels as a battery tender for a generator. It only is about 30w, but keeps a small 12v battery from dying over the months. It also has a charge controller built in - https://www.amazon.com/OYMSAE-Portable-maintainer-Cigarette-...
I suggest a different reasoning: what's about domestic p.v. with storage and racks in the basement with "free" A/C in terms of WFH and distributed "datacenters"?
How many have realized how much stuff can be hosted at home with availability levels not really far from most common datacenters?
Nicely done!
Now, Low Tech Magazine also has instructions to convert a stationary exercise bike into a human powered generator, which you could build to add power during the winter :-)
Cool project, but also a great all around tutorial - the appendix section is a nice idea and very helpful for many people I'm sure.
this is really neat.
This is cool, but I mean... come on. This guy lives in a $2M apt in a big city.
The amount of energy being "saved" yearly is wasted almost every second.
The author clearly explains in the post:
> When I started this solar-powered website project, I wasn't trying to revolutionize sustainable computing or drastically cut my electricity bill. I was driven by curiosity, a desire to have fun, and a hope that my journey might inspire others to explore local-first or solar-powered hosting.
> The cost savings? Looking at our last electricity bill, we pay an average of $0.325 per kWh in Boston. This means the savings amount to $2.85 USD per year (8.76 kWh * $0.325/kWh = $2.85). Not exactly something to write home about.
I can't speak for his specific apartment, but apartments in general are more energy efficient that a suburban house. New York City is one of the lowest per capita CO2 emissions in the US.
I can believe it either way, but some of that is because carbon-intensive industry will be based out of the city and its products imported in.
I doubt there’s a cement kiln, cattle farm or oil refinery within city limits.
These figures are generally based on consumption rather than production for that reason. It’s not helpful to credit emissions to the people who largely aren’t driving the output.
City living brings efficiencies from operating at scale - fewer miles by car, block of apartments instead of single family homes.
How does it help the discourse to bring this up? Would you like it if someone inserted this kind of (irrelevant) and unsubstantiated information when commenting on an article you wrote?