A scientific research institute needs to offload a large collection of genomic data sets from its on-premises servers to AWS. The data sets are seldom accessed, but when they are, a delay of several hours is acceptable. The institute requires a highly cost-effective solution for storing and retrieving these data sets, with a strong focus on minimizing storage costs. What method represents the MOST cost-optimized approach to store this data?
Utilize the Hadoop Distributed File System (HDFS) on Amazon EMR with occasional syncing to S3 for the data sets not actively in use.
Store the genomic data sets using the S3 Glacier Deep Archive storage class after initial upload completion.
Leverage S3 Intelligent-Tiering to automatically optimize costs between frequent and infrequent access tiers for the data sets.
Implement a Storage Gateway with stored volumes to gradually move the data sets into Amazon S3 over a direct connection.