As we, as a server collective irational.org, are making a low-energy
cost backup server based on a Raspberry pi. We are receiving advice from
the person who worked on Low-Tech Magazine’s solar server. We are
planning to install this at servus.at in Austria (already came to an
Since the energy crisis, there seems to be a larger demand for
single-board computers, I was very to find from an official vendor,
otherwise would have to go look for one on eBay where currently pis are
being sold for premium prices.
Maybe it would interesting to come up with THE recipe for a
low-cost/low-energy server. There are many SBCs out there and it would
be nice if anyone with a similar idea can avoid all the hustle that I
So my interest lies in
Removing functionality: serene minimalism to save energy
I wish Tsvetan could attend so that we could discuss such a setup on @olimex machines.
I can imagine a WOL setup to wake up the board when backups come in. I’m eager to see how you made it work. I remember that the baseline for Irational.org was to use standard Operating System with no customization in order to be able to adapt to any *nix computer at hand.
There are plenty of SBCs around nowadays, especially with the cm4 form factor there are nice baseboards available.
One interesting aspect I would like to discuss is: how do we do distributed offsite backups for friendly entities. E.g. what is there to make “2 copies in my trusted friend circle” + 2 copies wherever possible. And how does graceful on and offboarding work. (I share how much bandwidth? How much storage? Can I help locally by dropping a USB stick to speed up transfer? Can I click “bye” and after X weeks it is save to turn off sbc?)
So what are sane requirements that entities like hackerspaces could contribute, which might be cost sensitive. Can we share collections of media?
this is something that I would be interested in. I see some overlap and some more random stuff.
I’m very interested in this topic, as I’ve been tackling it from the software side. So, although I probably won’t be able to contribute much to the discussion when it comes to the hardware, I might share some insight as to what is possible in terms of available technology.
My main use case is keeping peer-to-peer data available, and even highly available (multiple redundant copies).
It doesn’t require much resources: ~150Mb RAM when idle, more when downloading/uploading large files, although the underlying stack should adapt if not much RAM is available–can investigate if the hardware you have in mind has very little ram
I’m using the hypercore stack, because they already solve a lot of the issues (file-level abstraction with Hyperdrive, encryption on transport and possible encryption at rest, sparse replication, offline first, and load-balancing downloads between multiple peers).