3 ms·
You're just adding more points of failure.
by mike-cardwell 13y ago
You're just adding more points of failure.
- pyre 13y agoI thought that a single point of failure was a bad thing. Sure you might want to only have a single place where things can fail, but if that failure is catastrophic...
- Drakim 13y agoNo, adding more points of failure doesn't make the first potential point of failure more fail-proof, or anything like that.
- carbocation 13y agoIf adding a new layer (gmail) creates a new point of failure without creating any new resilience against failure, you're putting yourself into a bad position. Think about failure like resistance. Points of failure in parallel gives less overall risk of failure. Points of failure in serial (like adding a front-end) gives greater overall risk of failure.
- nknighthb 13y ago"Single point of failure" means a point at which a failure effectively breaks the entire system or workflow. A system may have many SPOFs, which is worse than a single SPOF, which is, in turn, worse than no SPOFs (that is, the desired case is no single failure kills forward progress).
- squidi 13y agoYes, extra points of failure may include: - Having webmail which is not protected with two factor authentication - Enabling POP/IMAP which also does not have two factor authentication
- pyre 13y agoThis is true, but sometimes be introducing a new risk (point of failure) to the system, you get the ability to route around another. In this instance, separating POP and Gmail gives you the ability to route around Gmail when it's down.
- justincormack 13y agoNot really if you can also use the other account directly if Gmail were to fail.