There's also the hybrids. Only storing "unlikely to be accessed" data on S3 would likely remove a huge part of the AWS bill, while significantly limiting the si

There's also the hybrids. Only storing "unlikely to be accessed" data on S3 would likely remove a huge part of the AWS bill, while significantly limiting the size of the on-prem data set. Of all the possible related problems, object storage is the easiest.