Architectures to provide high availability, reliability, and scalability

We have come a long way in terms of making our systems more reliable, scalable, and available. It wasn't that long ago that we didn't think anything of saving precious photographs and documents on our PC hard drives, assuming that they were going to be able to store this data indefinitely. In reality, even though PC components have decent reliability, they will eventually fail. It's the nature of hardware with moving parts such as disk drives.

Since then, great advances have been made to increase the reliability of individual components, however, the real increase in reliability comes from redundantly storing information on multiple devices as well as in different locations. Doing so increases reliability exponentially.

For example, the S3 Standard service stores files redundantly with at least 6 copies and in at least 3 data centers. If a copy is corrupted, the S3 storage system automatically...