AWS Snowball Edge is no longer available to new customers. New customers should explore [AWS DataSync](https://aws.amazon.com/datasync/) for online transfers, [AWS Data Transfer Terminal](https://aws.amazon.com/data-transfer-terminal/) for secure physical transfers, or AWS Partner solutions. For edge computing, explore [AWS Outposts](https://aws.amazon.com/outposts/). 

# Quotas for compute instances on a Snowball Edge device
<a name="ec2-edge-limits"></a>

The following are storage quotas and shared resource limitations for compute resources on an AWS Snowball Edge device.

## Compute resource storage quotas on Snowball Edge
<a name="device-storage-limits"></a>

The storage available for compute resources is a separate resource from the dedicated Amazon S3 storage on a Snowball Edge device. The quotas for storage are as follows: 

A Snowball Edge compute-optimized device can run up to 20 AMIs and 10 volumes per instance.


****  

| Instance type | vCPU cores | Memory (GiB) | Supported device option | 
| --- | --- | --- | --- | 
| sbe1.small | 1 | 1 | storage optimized | 
| sbe1.medium | 1 | 2 | storage optimized | 
| sbe1.large | 2 | 4 | storage optimized | 
| sbe1.xlarge | 4 | 8 | storage optimized | 
| sbe1.2xlarge | 8 | 16 | storage optimized | 
| sbe1.4xlarge | 16 | 32 | storage optimized | 
| sbe1.6xlarge | 24 | 32 | storage optimized | 
| sbe-c.small | 1 | 2 |  compute optimized  | 
| sbe-c.medium | 1 | 4 |  compute optimized  | 
| sbe-c.large | 2 | 8 |  compute optimized  | 
| sbe-c.xlarge | 4 | 16 |  compute optimized  | 
| sbe-c.2xlarge | 8 | 32 |  compute optimized  | 
| sbe-c.4xlarge | 16 | 64 |  compute optimized  | 
| sbe-c.8xlarge | 32 | 128 |  compute optimized  | 
| sbe-c.12xlarge | 48 | 192 |  compute optimized  | 
| sbe-c.16xlarge | 64 | 256 |  compute optimized  | 
| sbe-c.24xlarge | 96 | 384 |  compute optimized  | 

## Shared compute resource limitations on Snowball Edge
<a name="shared-resource-limitations"></a>

All services on a Snowball Edge device use some of the finite resources on the device. A Snowball Edge device with its available compute resources maximized can't launch new compute resources. For example, if you try to start the NFS interface while also running a `sbe1.4xlarge` compute instance on a storage optimized device, the NFS interface service doesn't start. The following outlines the available resources on the different device options as well as resource requirements for each service.
+ If no compute services are `ACTIVE`:
  + On a storage optimized option, you have 24 vCPUs and 32 GiB of memory for your compute instances.
  + On a compute optimized option, you have 104 vCPUs and 208 GiB of memory for your compute instances.
+ While AWS IoT Greengrass and AWS Lambda powered by AWS IoT Greengrass are `ACTIVE`:
  + On a storage optimized option, these services use 4 vCPU cores and 8 GiB of memory.
  + On a compute optimized option, these services use 1 vCPU core and 1 GiB of memory.
  + While the NFS interface is `ACTIVE`, it uses 8 vCPU cores and 16 GiB of memory on a Snowball Edge device.
  + While Amazon S3 compatible storage on Snowball Edge is ACTIVE on a Snowball Edge Compute Optimized with AMD EPYC Gen2 and NVME, for a single node with the minimum configuration of 3 TB of Amazon S3 compatible storage on Snowball Edge, it uses 8 vCPU cores and 16 GB of memory. For a single node with more than 3 TB of Amazon S3 compatible storage on Snowball Edge, it uses 20 vCPU cores and 40 GB of memory. For a cluster, it uses 20 vCPU cores and 40 GB of memory.

You can determine whether a service is `ACTIVE` on a Snowball Edge by using the command `snowballEdge describe-service` on the Snowball Edge client. For more information, see [Vieiwing status of services running on Snowball Edge](using-client-commands.md#client-service-status).