Does amateur/semi-professional self-hosting have a real future?
It’s been three months now that I’ve been “playing around” with a VPS to see what I could do with it. I already did all this stuff 15–20 years ago. So I tried the usual stuff: email, calendar/contacts, rclone, Firefox Sync, etc...
But the more I’ve progressed, the more I’ve scaled back my need for “device-to-device” solutions, or considered prioritizing “relay” systems over traditional servers (a Syncthing relay rather than a server accessible via rclone).
But even recently, after reading so much about security, it’s become very difficult to consider.
While 20 years ago, my limited knowledge allowed me to handle simple setups, it seems that’s no longer possible today.
I see your posts about securing your configurations, and it’s becoming very complex and a constant source of stress.
I’ve always hoped that encouraging people to have their own personal servers would go hand in hand with making them easier to use.
Unfortunately, that is not the case.
With a few exceptions, we should be able to avoid all of this and manage our data from our client devices, and between clients, without relying on servers anymore—except for an off-site backup of our data, encrypted end-to-end. But we should no longer have to let a remote machine manage and process sensitive personal data in plain text and under the control of a system administrator (prefer a system like KeePass over Vaultwarden, for example).
We should rely solely on audited open-source software, including only device-to-device synchronization systems. All processing should take place on the clients, with only archiving handled on the servers. Obviously, this is trickier if you manage online services, but some people find a compromise using relays/instances with Nostr/Mastodon, though that isn’t always possible.
What do you think?
P.S.: I wrote this all myself, starting from the French version, using Deepl’s translation.
P.S. 2:
I’ve used examples to illustrate a feeling and a broader issue. Please don’t reduce my questions to ad-hoc technical answers based on specific examples. That would only serve to highlight the reality of the problem I’m raising: the complexity we trap ourselves in.