A company ingests millions of 1 MB image files per day and analyzes them in 1 GB batches. After analysis, the application zips each batch and archives the zip file to an on-premises NFS server. The company has Hyper-V compute capacity on premises but no additional storage. It wants to archive images on AWS, be able to retrieve archived data within one week, and schedule transfers during non-business hours while enforcing bandwidth limits over a 10 Gbps Direct Connect link. Which cost-effective solution meets these requirements?

The correct solution uses an AWS DataSync agent deployed as a Hyper-V VM on premises, which is cost-effective as it leverages existing Hyper-V infrastructure without needing new EC2 instances. DataSync is ideal for large-scale, scheduled data transfers with bandwidth control over Direct Connect. Archiving to Amazon S3 Glacier Deep Archive is the most cost-effective storage tier for data that needs to be retrieved within one week (retrieval times can be as fast as 12 hours) and is accessed infrequently, aligning with the archiving requirement. Deleting on-premises data after successful transfer frees up local storage. The incorrect options are: Deploying a GPU-based EC2 instance for DataSync is unnecessary and expensive for this use case. S3 Glacier Instant Retrieval is more expensive than Deep Archive and not required given the one-week retrieval window. Deploying a general-purpose EC2 instance for DataSync adds unnecessary cost, and transitioning to S3 Glacier Deep Archive after 1 day via lifecycle rules is less direct than archiving directly. Storage Gateway Tape Gateway is designed for virtual tape libraries and backup applications, not for direct file archiving from an NFS share, and managing virtual tapes adds complexity not required here.