Not to sound like a wet blanket, but I think you’ll find it hard to make enough money to cancel out your power bill, development and admin time compared to big box names like AWS, digital ocean, etc. It could also open you up to legal problems. If someone you don’t know wants to pay you (someone with no reputation they shouldn’t trust with their money and data) then you probably don’t want the data they are trying to host any where near your servers (use your imagination…).

That said, if you’re serious or want a learning exercise look for:

• identity management/authentication - SSO or Oauth for an authentication layer. examples: Keycloak, Oauth proxy.
• resource monitoring - Metrics, monitoring, logging. some of this is just for maintaining availability of your service but you may need metrics for usage billing or cutting off access if they go over allotments. Examples: ELK stack, Prometheus, grafana, Loki, datadog
• API gateway - api rate limitingbased on identity, e.g. user Bob is only allowed to make 1000 API calls a month . Examples: 3scale.
• domain specific resource access policies - this is where it gets tricky. Depending on what application you are offering as a service, there may be access/limitation policies built in you could use or generic external options. If not, you’ll need to come up with a method on your own like an authorization proxy that intercepts requests and approves or denies based on identity and usage metrics. A decent API gateway solution should allow for this customization.

I am not aware of an out of the box solutions that solve all of these problems, but there are many solutions for specific parts of this hypothetical system which is what I tried to provide with some example tech names to start your search. It’s not a simple problem, but it’s one basically every internet company has solved on some level.

I think there are croud-source compute reselling services you can enroll your systems in but I’m blanking on names and again you’ll be lucky to break even on power.