Using Amazon S3 compatible storage on Snowball Edge - AWS Snowball Edge Developer Guide

Using Amazon S3 compatible storage on Snowball Edge

Amazon S3 compatible storage on Snowball Edge delivers secure object storage with increased resiliency, scale, and an expanded Amazon S3 API feature-set to rugged, mobile edge, and disconnected environments. Using Amazon S3 compatible storage on Snowball Edge, you can store data and run highly available applications on Snowball Edge for edge computing.

You can create Amazon S3 buckets on the Snowball Edge devices to store and retrieve objects on premises for applications that require local data access, local data processing, and data residency. Amazon S3 compatible storage on Snowball Edge provides a new storage class, SNOW, which uses the Amazon S3 APIs, and is designed to store data durably and redundantly across multiple Snowball Edge devices. You can use the same APIs and features on Snowball Edge buckets that you do on Amazon S3, including bucket lifecycle policies, encryption, and tagging. When the device or devices are returned to AWS, all data created or stored in Amazon S3 compatible storage on Snowball Edge is erased. For more information, see Local Compute and Storage Only Jobs.

You can deploy Amazon S3 compatible storage on Snowball Edge in standalone configuration or in cluster configuration. In standalone configuration, you can provision S3 capacity on the device and the balance is available as block storage. In cluster configuration, all data disk capacity is used for S3 storage. A cluster may consist of a minimum of 3 devices up to a maximum of 16 devices. Depending on the size of cluster, S3 service is designed to sustain device fault tolerance of 1 or 2 devices.

With AWS DataSync, you can transfer objects between Amazon S3 compatible storage on Snowball Edge on a Snowball Edge device and AWS storage services. For more information, see Configuring transfers with S3 compatible storage on Snowball Edge in the AWS DataSync User Guide.

Following is the Amazon S3 compatible storage on Snowball Edge storage capacity and block storage capacity for a standalone device using Amazon S3 compatible storage on Snowball Edge. For fault tolerance and storage capacity of clusters, see this table.

Snowball Edge Compute Optimized with NVMe storage
Storage capacity of Amazon S3 compatible storage on Snowball Edge and block storage of Snowball Edge Compute Optimized (Compute Optimized with AMD EPYC Gen2 and NVMe) devices
Amazon S3 compatible storage on Snowball Edge storage capacity (in TB) Block storage capacity (in TB)
3 17.5
5.5 14.5
10.5 8.5
12 6.5
13 5.5
16.5 1.5
Snowball Edge storage optimized 210 TB
Storage capacity of Amazon S3 compatible storage on Snowball Edge and block storage of Snowball Edge storage optimized 210 TB devices
Amazon S3 compatible storage on Snowball Edge storage capacity (in TB) Block storage capacity (in TB)
20 206
40 182
60 158
80 134
100 110
120 86
140 62
160 38
180 14
190 2

Amazon S3 compatible storage on Snowball Edge specifications:

  • The maximum number of Snowball Edge buckets is 100 per device or per cluster.

  • The S3 on Snowball Edge bucket owner account owns all objects in the bucket.

  • Only the S3 on Snowball Edge bucket owner account can perform operations on the bucket.

  • Object size limitations are consistent with those in Amazon S3.

  • All objects stored on S3 on Snowball Edge have SNOW as the storage class.

  • By default, all objects stored in the SNOW storage class are stored using server-side encryption with Amazon S3 managed encryption keys (SSE-S3). You can also explicitly choose to store objects by using server-side encryption with customer-provided encryption keys (SSE-C).

  • If there is not enough space to store an object on your Snowball Edge, the API returns an insufficient capacity exception (ICE).

Topics