What is a term for ineffective security measures that don’t prevent any realistic attack?
Is there a term for when you a particular system design might prove to have some advantages, but doesn't actually qualitatively change the potential attacks on the system and thus ends up as redundant, excessive, misleading, costly, and/or ineffective?
For example, a programmer might say, "We should make the user type in TWO passwords upon every login. We'll store one password in one database and the second password in a different database."
Many would deem a second password as unnecessary, as you could instead just increase password length or complexity. And it doesn't matter if the system allows unlimited password attempts or contains similar vulnerabilities. It also begs the question of "well, why not THREE password then?"
Although, having two passwords in separate databases certainly adds some sort of extra hurdle for an attacker, especially because a single database dump doesn't give you all sensitive information. But my impression is that, in practice, it is more secure to have a single source of truth for sensitive data that is as clear as possible, and then protect that data as well as possible. This is like the idea behind using a bastion host for SSH access: it turns out it is more secure to handle things in one spot, rather than leaving it to each host to secure itself, even though there may be a case where an attacker fails to gain access to a certain host having a coincidentally different SSH configuration.
