I wanted to know if there is an option to use Amazon S3 as a database by taking advantage of Amazon CloudSearch to be used as a query-like engine. For example, to make paging, I might use Boolean expression and use letters as incremental values. Inserting and updating data can be done easily.
Also backup and restore can be done easily by backing up the bucket. You don't have any size storage limitations. You can use caching in order to scale the application and even use CDN together with some headers to scale.
cloudSearch can be a great query-like engine especially for search (it was designed for that of course), and it can scale out without any need for maintenance.
I wanted to know if it can be made? Also taking into account that we can use map/reduce (if I remember right).