We decided to disaggregate the system and user data storage, but our current implementation still reads data from the system storage in DynamoDB. - [ ] Read data from S3 buckets. - [ ] Decoding of bytes data - user data and properties. - [ ] Document strong consistency requirements.