#selfhost #homelab #opensource #foss
Hello dear Fediversers,
My students want to improve their home lab backups and have come up with the idea of storing their offsite backups with each other instead of in the cloud.
My question is, does anyone know of any software or solutions for this? It doesn't have to be open source, but it would be best if it were. The option to encrypt the backups is not essential, but would be preferable.
Please boost.
Many thanks for any tips.
They can upload them to a decentralised private cloud.
IPFS is not the best option IMO because it requires ongoing payments. But Autonomi requires only a one off payment and is encrypted.
My blog explains how this works for publishing websites using Autonomi Dweb, but websites are just files and the process for raw files is similar. But using Dweb you can provide a UI for downloads as a website if needed.
See: https://toast.happybeing.com
Hope that's of interest!
@jwildeboer @kkarhan
@happyborg @dalai @jwildeboer I'd not trust any "one time purchase" cloud storage because that model is unsustainable by virtue of recurring costs!
If one's already holding classes for #IT, it's either feasible to organize some (quota'd!) storage (i.e. #SSH access to a server) or expect the class members to have i.e. some cheap external HDD for that.
I don't expect them to backup several terabytes of data and if they need to, then there should certainly be budget for either S3 block storage and a cheap VPS or a "Storage VPS"…
Not to mention there's need to not just backup files, but databases.
https://en.wikipedia.org/wiki/InterPlanetary_File_System
https://en.wikipedia.org/wiki/BitTorrent
@kkarhan
> I'd not trust any "one time purchase" cloud storage because that model is unsustainable by virtue of recurring costs!
You may not trust it but you haven't I believe looked into this case. In which case your judgement is premature.
I won't argue against your intuition or gut feel for this, which I accept as understandable. Nor that I know that you are wrong.
Having understood the arguments for this, I believe it needs testing because it is a thing worth having.
@happyborg @dalai @jwildeboer most of these providers claim that they can recoup the cost with new purchases and shrinking costs for storage over time, which makes this truly a #PyramidScheme.
OFC they also throtthe users and don't expect everyone to max out their purchased storage quota instantly.
But I'm not just comparing against short-term filehosters but also in terms of #backups consider years, if not decades of uptime to be necessary to be useful.
Also mind you this ain't like #TeamViewer where the cost of infrastructure is negligible (just some "Rendrevous-Server" to exchange status info, IP addresses and facilitate "Hole Pubching" through #NAT & #Firewall which in theory doesn't reuire mich compute and bandwith to function.
We're talking Gigabytes if not Terabytes per User in storage and traffic.
And since this system isn't like some #ColdStorage like a #tape, this needs to be #HDDs that constantly spin and draw power.
@kkarhan
It's a valid approach and until tested it is premature to say it doesn't work.
You're entitled to be skeptical. I'm not convinced either - but there are plenty of things that intuitively I believe "can't work" that I see with my own eyes do in fact work.
My point is not that you are wrong, but that you do not know that you are right.
Your argument is emotive rather than rational IMO.
@dalai @jwildeboer
@happyborg @dalai @jwildeboer Well, unless I extremely overcharge clients for an indefinite amount of time, I'd either have to shutdown after some time or intentionally banrupt at some point.
The cost of storage ain't shrinking (currently) and whencthey do, it's a matter of years.
There are only a finite amount of people able to pay such a sum upfront and unless the revenue is being diverted into something more profitable (which would likely run afoul with accounting laws) the chances of this not bursting are proportional to the amount of storage and speeds offered...
Even if you discount all backups and redundancies and only calculate with consumer HDD pricing you'd have problems sooner or later
Unless that price would far exceed even the most aggressive, avcident-prone and fuel-wasting driver, it would not make sense and certainly if it means said company willcat some point be making partd for decades...
@kkarhan
You make some valid points to back up what are opinions and predictions, and some obviously questionable. But you don't know the future and pay no attention to mitigations or alternate scenarios. Only a closed mind is certain.
If the model fails, it can be adjusted. But ruling things out without trying is not how humans have become such incredible innovators, so I prefer to try and solve problems that can create desirable outcomes.
Dalai
happyborg
Kevin Karhan 
