AWS Snowball Edge

Device Configurations

Snowball Edge devices come in two main configurations, tailored for different use cases.

1. Snowball Edge Storage Optimized

This device is designed for large-scale data migration and capacity-oriented edge workloads.

• Storage:

  • Usable S3-compatible storage: 80 TB HDD

  • Usable Block storage: 80 TB HDD

  • NFS storage: 80 TB HDD

  • SSD storage for cache: 1 TB NVMe

• Compute:

  • vCPUs: 40

  • Memory: 80 GB RAM

• Best For: Large data transfers, data consolidation, and workloads requiring high storage capacity at the edge.

2. Snowball Edge Compute Optimized

This device provides powerful computing resources for machine learning, data analytics, and other processing-intensive workloads at the edge.

• Storage:

  • Usable S3-compatible storage: 42 TB HDD

  • Usable Block storage: 42 TB HDD

  • SSD storage for performance: 7.68 TB NVMe

• Compute:

  • vCPUs: 52

  • Memory: 208 GB RAM

  • Optional GPU: Available with an NVIDIA Tesla V100 GPU for advanced machine learning and data analysis.

• Best For: Edge machine learning inference, video analytics, running compute-intensive applications in disconnected environments.

How It Works: The Data Transfer Workflow

1. Plan and Create Job: In the AWS Management Console, you create a job, specifying the job type (e.g., Import to S3), your shipping address, and the Snowball Edge device configuration you need.

2. Receive the Device: AWS ships the device to your location. The device is rugged and comes in its own shipping container.

3. Connect and Unlock: You connect the device to your local network and power it on. To unlock it, you use the AWS OpsHub for Snow Family application (a GUI tool) or the AWS CLI, along with a manifest file and a 25-character unlock code provided in the console.

4. Transfer Data:

   • Use the S3-compatible endpoint on the device to transfer object data using the AWS CLI or SDKs.

   • Use the NFS file interface for file-based data transfers.

   • Launch EC2 instances or Lambda functions for edge processing tasks.

5. Ship Back to AWS: Once your data transfer or edge job is complete, you power off the device. The E Ink shipping label automatically updates with the correct AWS return address. You then ship the device back to AWS via the designated carrier.

6. AWS Completes the Job: Upon arrival at an AWS facility, the data from the device is imported into your specified S3 bucket. You can track the entire process via the console and Amazon SNS notifications.

Key Features

• Edge Compute: Run your own applications on Amazon EC2 instances or using AWS Lambda functions directly on the device, even when disconnected from the internet.

• S3-Compatible Endpoint: The device provides an S3-compatible object storage endpoint, allowing you to use existing S3 APIs and tools to transfer data.

• NFS File Interface: Offers a simple file-based interface for easy data transfer from file-based systems.

• Clustering: You can cluster 5 to 10 Snowball Edge devices together to create a single, larger, and more resilient storage pool with local data durability.

• AWS OpsHub: A graphical user interface you can install on a local computer to manage the device, transfer data, and launch compute instances without needing to be an expert with the AWS CLI.

Security

Security is a core design principle of the Snowball Edge.

• Rugged Enclosure: The device is designed to be tamper-evident and withstand harsh environmental conditions.

• Encryption:

  • All data written to the device is automatically protected with 256-bit encryption.

  • You manage the encryption keys using the AWS Key Management Service (KMS).

• Trusted Platform Module (TPM): The device has a dedicated TPM chip that provides a hardware root of trust, protecting against unauthorized modifications.

• Secure Erasure: After your data is successfully imported into S3, AWS performs a software erasure of the device that follows NIST guidelines for media sanitation.