One day I stumbled upon a list on GitHub Kickball/awesome-selfhosted listing open-source services and applications that can be hosted locally or on my own servers. The list covers much more than you’d expect, you can find alternatives to most close-sourced services you use daily on the list. For example, Nextcloud does mostly what G Suite does; use GitLab to replace GitHub; use Mastodon, the free (as in freedom) clone of Twitter…
Nowadays high-quality (and some not-so-high-quality) online services are available for free to the public in every steps of life. All it takes is usually just an email address and a password, maybe phone verification too, then one can enjoy limitless photo backup on Google Photo, reach millions of people on Twitter, etc. Sounds awesome? No, there comes the hidden cost as you hand over your data and willingly subject yourself to their experiments and selection of ads.
Of course, I don’t mean stop using all those awesome free services and start self-hosting everything yourself. That’ll be impractical and prohibitively expensive in both storage and actually running the services. The good thing is, there are always self-hosted alternatives to opt for.
As it turned out, it wasn’t a easy task to self-host anything, not at all.
Hosting my own note-taking service
The time was the beginning of the semester, when I had to decide which note-taking application to use for the semester. I used to take note with pen and paper. They are ineffective and too heavy to carry around near the end of the semester.
I needed something that syncs and backups automatically, works on both laptop and mobile, and should look nice enough. Originally it wasn’t about self-hosting. I took notes on Vim and sync them to GitLab. Shortly after, I trashed this practice because my notes are always out of sync and I couldn’t read them on the phone.
Then I found Standard Notes.
Standard Notes is just one app thats connects to a Standard File server. Think of it as a personal database that can be used across all apps in an encrypted, standardized format. With Standard File server, you own your own data. Nobody can tamper with your data. Nobody can look into the data and tell what kind of ads or site you should visit. Standard File server should remain private and encrypted, end-to-end. As such, I was sold.
The time is 2017, it is easy to gather the necessities to self-host something. Domain name? GoDaddy. SSL? Let’s Encrypt or Cloudflare. Hosting? Digital Ocean/Linode. With the proper knowledge and choice of service provider it shouldn’t break your bank trying to self-host something.
However there were a few issues. Setting up a Standard File server could be a lengthy process involving a bunch of command crunchings. To facilitate the process one could use Docker, if there was a Dockerfile with documentation.
Turned out there was no proper Dockerfile merged to the repo
standardfile/ruby-server
and no documentation regardling deployment using Docker whatsoever. It
took me two nights to modify a Dockerfile from a fork, and a few more nights to finish a
proper one and get it merged.
Cool, now can I finally deploy it and don’t have to worry about it for the decades to come? Not really.
When you own a server out there you’d have to worry about what any sysadmin would worry about. You should backup regularily, prevent attacks, keep the software up to date, monitor the uptime and remember to pay for the domain and hosting. This may seem to be a no-brainer for seasoned admins but as someone new to the field, it felt frustrating to configure multiple additional software just to get one self-hosted online service running. A couple hours should be enough to configure everything and you can eventually begin to enjoy your own self-hosted service that probably won’t work better then its commercial counterpart, but it’s free, and you own the data.
Very soon you may think it’s too crappy and want to modify the source code yourself. That might take a couple more days away from your life albeit in the open-source community perspective it’s the preferable way to do things.
So, should I start self-hosting?
Nobody has to self-host something if free commercial services should suffice. But one big feature of those services is the safety of your data, even if the company hosting it is sold. Government, good or evil, could forcefully access your data. If you believe your data worths more protection then their own business interest, think again, think thrice if you have to. You may not have something to hide, most people don’t. That doesn’t mean in essence someone can legitimately look at your data without asking.
Self-hosting gives you the option to lock something up and throw away the key when necessary. While it is possible to reset password on most online services, you can keep the password that also serves as the encryption key of your data only in your head. If you decide to hit yourself in the head to forget the password, nothing better than brute-force guessing can recover the data.
You pay your hosting company to keep the server running every month. For anything that you’d rather take to your grave, just leave it. Take Digital Ocean as an example, your data will be purged if you don’t pay the bill for a long enough time.
Self-hosting is difficult to set up and troublesome to maintain. Most of the the open-source self-hosting solutions are less user-friendly and less functional then their commercial counterparts. Domain name renewal and hosting may cost a lot in long run. In exchange, you can own your data and peace in mind knowing nobody can sell your data for profit or access it without your authorization (unless they hack, but who cares about your data).