I looking after cheap databases for when I make new projects. The are often small projects with no intention often to make money, BUT I really don't want to host a database myself. The reasons is many, but most importantly backups! You have to extremely careful to watch database cron jobs to be sure the aren't failing.<p>The definition on cheap can be many, but my idea is that it only cost $1 per gigabyte stored or less and that I can only $1 dollar from the start.<p>I know it is extremely cheap, but there is already a lot of cheap PHP hosting providers that provide hosting and database, but it is always mysql and nothing else and there "limit" on the database is always unclear (connections, size, table count, user count, etc.) Heroku provides postgres for free, but only 10k rows and then goes directly to $9 dollars.<p>GCloud new SQL is awesome, and it only cost $7,5 + storage.<p>I know there are some NoSQL providers that gives 0,5G for free, but I am not interested in nosql, but why aren't there similar providers for SQL?<p>Am I one of a kind to want this?
What's wrong with SQLite?
Backup is a cron job to move a file.
Keep the db in the same place as the site if it is low value.
No expense but the expense you already have made for whatever server is hosting the content that isnt the database.
Basically every OS has drivers for it, and it has really good performance for such a small package (especially if you dont have oodles and oodles of data)
What about Amazon RDS? What kind of performance do you need?<p><a href="https://calculator.s3.amazonaws.com/index.html#s=RDS" rel="nofollow">https://calculator.s3.amazonaws.com/index.html#s=RDS</a><p>Indicates you can get 5GB for $0.00 as long as you don't want performance.<p>Disclaimer--I haven't done this.
A single google cloud sql baby instance should host 10, 20, 50 baby projects with no problem! Just use a different schema per project and you can split it later if you are successful.
You can strap on multiple low end yet cheap VPS servers with Redis. It can deliver decent performance but harder to keep security and reliability together, or if you want SQL go for Mongo.