DataSync vs S3 File Gateway
AWS DataSync and Amazon S3 File Gateway both sit between on-premises storage and Amazon S3, which is why they get confused. The difference is that DataSync runs a transfer and stops, while File Gateway is a permanent piece of infrastructure that applications keep talking to.
Key differences
Section titled “Key differences”| DataSync | S3 File Gateway | |
|---|---|---|
| Purpose | Migration and scheduled synchronisation | Ongoing storage access |
| Data movement | Point-in-time, per task run | Continuous, as applications read and write |
| Local cache | None | Caches frequently accessed data locally |
| Interface | Task configuration, no mount point | NFS or SMB mount point |
| Charging | Per gigabyte transferred | Storage, requests and gateway usage |
| Targets | S3, EFS and all four FSx file systems | S3 |
| Lifespan | Deployed for the duration of a project | Permanent infrastructure component |
Use DataSync when
Section titled “Use DataSync when”- Performing a large-scale data migration
- Scheduling recurring transfers or replication
- Moving data between AWS storage services, or in from another cloud
- Automated validation and verification of the copy matters
- Detailed transfer metrics and logging are needed
Use S3 File Gateway when
Section titled “Use S3 File Gateway when”- Applications need continuous access to data held in S3
- Local caching is needed for acceptable read latency
- The application must see a file share and cannot be changed to call the S3 API
- NFS or SMB is a hard requirement
- The hybrid architecture is the end state, not a migration step
Using both
Section titled “Using both”The two compose well. Run DataSync once to move the bulk of an existing share into S3, then put a File Gateway in front of the bucket so the on-premises applications keep working against what looks like the same share. The migration finishes quickly and the operational access pattern survives it.