Use S3 transfer acceleration

[sanderegg]

This is the reference: https://aws.amazon.com/s3/transfer-acceleration/

This should be:

• tested
  • check pricing
  • create a new bucket set it accelerated
  • copy a large dataset (TN,300k files and 170Gib)
  • start EC2, rclone pull/upload and measure timing
  • compare to standard bucket
  • do it in us-east-1/us-east2
• used if it improves the performances