We are exploring how SQL inject risks can be limited by only allowing key functions to perform duties within our rails application. I am the security officier, not a full time rails person.
The way I see it, only writing to the database should occur when an account is created (registration) or the account is modified (user profile update). For simplification purposes, let's assume that no other updates happen within the application from users.
If someone does not block a SQL injection attack on a read only request, we would be protected better with two separate calls.
Can rails support the following configuration? -
- One database connection for all "read" requests
- One database connection (same exact database) for all "write" requests
If so, how would that look in the database file and code requests?