Hacker Newsnew | comments | show | ask | jobs | submit login

The analogy doesn't quite hold. The key idea behind a Redundant Array of Independent Disks is that under the assumption that disk failures occur independently with known probability, you can compute the probability of the array remaining intact.

In contrast, with a Redundant Array of Insolvent Banks, there is a very high correlation between bank failures -- but you're saved by the FDIC per-bank insurance limit. You're not modelling your risk based on known probability distributions; you're eliminating your risk entirely (at least as long as you trust the FDIC).




You're not modelling your risk based on known probability distributions; you're eliminating your risk entirely (at least as long as you trust the FDIC).

To (over?)extend the analogy further, there's always the possibility of a controller failure.

-----


...and each bank, ideally, guards against insolvency by distributing your deposits among a Redundant Array of Independent Debtors.

-----


or, he was just talking about RAID0

-----




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: