S3 Application Integration

Configuring existing Amazon S3 applications to work with Swarm consists of changing the region end-point and changing the authentication credentials.

Best Practice

Start with the documentation provided by Amazon Web Services and then use this section to help you integrate your S3 applications with the Swarm platform.

Within your S3 applications, change the following items:

Region End-Point – Use the Swarm storage domain name instead of the Amazon S3 region end-point hostname.
Access Key – In the Content UI, create an S3 authentication token in the correct domain and bucket, then use that token ID as your Access Key ID. For creating tokens in Content UI, see https://perifery.atlassian.net/wiki/spaces/public/pages/2443819264.
Secret Key – From the same token, use its secret key value in your S3 applications.
Update your configuration for best results:
- Enable "path style" access to avoid certificate validation failure; otherwise, the client/SDK may attempt to access as mybucket.mydomain.example.com/object.
- Use Version 2 signatures with the AWS .NET SDK for best performance. See https://docs.aws.amazon.com/general/latest/gr/signature-version-2.html.
- Increase the part size to 100 MB or more for multipart uploads, if configurable.

Warning

Old versions of the AWS S3 SDK for C# require that the “endpoint” contains “s3-” or “s3.” to avoid a 403 SignatureDoesNotMatch error when the object name has spaces or other special characters. The domain, or bucket if using “bucket in Host” style, must contain “s3” followed by a dot or dash. Upgrade to a recent C# SDK release (for example, 2021’s AWSSDK.S3 3.5.10.2 and AWSSDK.Core 3.5.3.9 or later) to avoid this requirement.