Skip to main content
In this lesson we’ll cover the AWS Snow Family — the suite of rugged, secure, physical devices and services for moving large volumes of data to and from AWS and for running edge compute workloads. Instead of pushing terabytes or petabytes over a WAN, you can write encrypted data to a Snow device, ship it to AWS, and have AWS import the data into your account. This approach can significantly reduce transfer time, lower network costs, and solve connectivity or bandwidth constraints. Why choose a physical transfer with the Snow Family?
  • Large datasets take a long time to upload over typical internet links and may incur substantial network transfer costs.
  • Physical devices avoid dependency on sustained bandwidth and often complete large migrations faster than network-based transfers.
  • Snow devices address security, reliability, and connectivity challenges in remote, edge, or harsh environments.
Before ordering a Snow device, create and configure your destination Amazon S3 bucket(s) and ensure the necessary IAM permissions are in place. Preparing the S3 destination and roles ahead of time simplifies import configuration and speeds up ingestion when the device arrives.
Features and benefits
  • Rugged and portable
    Snow devices are designed to tolerate shipping and field conditions. Snowcone is ultra-portable for constrained environments; Snowball devices look like a rugged suitcase for larger-capacity needs. These devices are well-suited for ships, construction sites, aircraft, and other deployments outside traditional datacenters.
  • File access / NFS mount
    Selected Snow devices support NFS (NFSv3 and NFSv4.1), allowing them to be mounted by on-premises servers and file-based applications. File metadata is preserved; when imported into S3, files are converted to objects.
  • Edge compute
    Compute-capable Snow devices can run selected Amazon EC2 instances and AWS IoT Greengrass functions. You can preload AMIs during ordering to run local processing, filtering, image/video analytics, or inference before shipping only the results or filtered datasets.
  • Strong encryption and hardware root of trust
    All data written to Snow devices is encrypted using 256-bit keys managed by AWS KMS. The device never stores the keys. Each device includes a Trusted Platform Module (TPM) to provide a hardware root of trust and tamper resistance.
  • Offline data transfer
    Devices are shipped with your encrypted data to a specified AWS Region, where AWS imports the data into the S3 bucket(s) you specified. The transfer itself does not require an internet connection.
  • Tamper-evident shipping and tracking
    Snow devices are tamper-evident and use an E Ink shipping label for automated tracking updates. Track job status and device delivery through the AWS Console, Amazon SNS notifications, and text messages. After AWS completes the import, the device is securely wiped before leaving AWS custody.
Device types
  • AWS Snowcone
    The smallest, ultra-portable Snow device for collecting, processing, and moving data in constrained environments. Snowcone supports both offline shipment and online transfer (via AWS DataSync) in environments with intermittent connectivity. It’s optimized for mobile field use and local edge compute.
  • AWS Snowball (including Snowball Edge family)
    Snowball devices are rugged, transportable appliances for bulk data transfer. Variants include:
    • Snowball (storage-focused): Primarily for straightforward offline data migration and storage-focused transfers.
    • Snowball Edge: A suitcase-sized appliance that adds onboard compute alongside storage. Snowball Edge comes in configurations optimized for compute or for storage so you can pick the right balance of local processing and capacity.
  • AWS Snowmobile
    An exabyte-scale data transfer service using a dedicated semi-trailer truck. Use Snowmobile to migrate extremely large datasets (petabyte- to exabyte-scale). An individual Snowmobile can handle up to around 100 PB per shipment (capacity and offerings may vary).
A slide titled "Snow Family – Components" showing icons and labels for AWS Snowball Edge, AWS Snowcone, AWS Snowball, and AWS Snowmobile. Below are colored circles highlighting device illustrations and specs like CPU optimized, storage capacities, and exabyte-scale data migration.
Comparing Snow devices (at-a-glance) Typical use cases
  • Disconnected, intermittent, or limited connectivity: Use Snowcone and Snowball Edge when persistent connections to AWS are not available.
  • Edge analytics and IoT: Run local analytics, ML inference, or video/image processing on Snowball Edge or Snowcone and ship only the essential results to the cloud.
  • Large-scale data migrations: Use Snowball or Snowmobile when migrating multiple terabytes to exabytes of on-prem data to S3 is impractical over the network.
  • Harsh or remote environments: Rugged construction and offline operation make Snow devices suitable for maritime, field research, military, and industrial scenarios.
Integrations and workflow
  • Destination: All Snow Family imports target Amazon S3. Configure S3 buckets and IAM roles/policies before placing an order.
  • On-device compute: Compute-capable Snow devices can run Amazon EC2 instances and AWS IoT Greengrass; some devices also support Lambda-style workloads depending on the device configuration.
  • Online transfer option: Use AWS DataSync with Snowcone for online transfer when intermittent connectivity is available.
  • Notifications and tracking: Job status and delivery updates integrate with Amazon SNS for notifications and can be tracked from the AWS Console.
Security reminder: Snow devices encrypt data with AWS KMS-managed 256-bit keys and include a hardware TPM. Ensure your KMS keys, IAM roles, and S3 bucket policies are configured correctly before the device arrives. AWS does not persist your encryption keys on the device.
Summary AWS Snow Family devices provide secure, rugged options to transport large datasets and run edge compute when network-based transfers are impractical. Choose based on:
  • portability and field constraints (Snowcone),
  • balance of capacity and local compute (Snowball / Snowball Edge),
  • extreme, truck-scale migrations (Snowmobile).
Links and references

Watch Video