Which technique is described for initial data transfer when data volumes are large, aside from ongoing replication?

• A. Use only DMS for initial load.
• B. Migrate using a big-bang approach.
• C. Use VPN-only transfer.
• D. Snowball/Storage Gateway for initial data transfer.

Answer: (D)

When moving very large data volumes to AWS, the fastest and most reliable way to seed the initial dataset is to transfer the data offline using a physical device or gateway. This approach avoids on-the-wire bandwidth limits and potential network contention, letting you load enormous amounts of data quickly and securely before starting ongoing replication.

Using a Snowball or Snowball Edge device lets you copy your data on-site, ship the device to AWS, and have the data securely uploaded into S3. These devices are designed for large-scale migrations, offering encryption, tamper-evident seals, and verification, and they scale to handle terabytes to petabytes as needed. Storage Gateway can also play a role in initial seeding by facilitating offline transfer paths and staging data that will be uploaded to S3, while still enabling ongoing data transfer once the initial load is complete.

Other methods fall short for bulk initial loads: relying solely on a database migration tool for the initial bulk can be slow and network-bound; a big-bang approach introduces significant risk and potential downtime; and transferring over VPN only compounds bandwidth and time constraints, especially when data volumes are extreme.