Thanks for the reply. Relying on bursting for a core service like a db would be too much risk - it would make sense for batch compute or similar workloads. Fro

Thanks for the reply. Relying on bursting for a core service like a db would be too much risk - it would make sense for batch compute or similar workloads. From a systems viewpoint, I can’t understand if there’s an actual technical challenge involved or if this pricing is just a lack of competitive pressure. Is there some corner case about the POSIX guarantees of EFS that makes higher throughput challenging to achieve? (Otherwise one could just spray file blocks across many servers and maintain a metadata mapping of file to blocks to achieve good throughput)