"doing this properly on EC2 is an order of magnitude harder than doing it on dedicated hardware; the assumption on EC2 has to be that things will fail and you h

"doing this properly on EC2 is an order of magnitude harder than doing it on dedicated hardware; the assumption on EC2 has to be that things will fail and you have to build accordingly" I don't follow, how is that different from dedicated hardware where you also have to assume failure will happen? If anything I'd argue it's the opposite (but ultimately note that I think both take some work). But I say that it's perhaps the opposite because on EC2, you have EBS: "volumes that operate with 20 GB or less of modified data since their most recent Amazon EBS snapshot can expect an annual failure rate (AFR) of between 0.1% – 0.5%, where failure refers to a complete loss of the volume. This compares with commodity hard disks that will typically fail with an AFR of around 4%" And neither failure rate involves RAID which you can of course do with both EBS (mdadm) and disks to go for lower AFR projections.