On a project involving Heroku, I was considering placing some services on Linode and accessing them from Heroku, which is hosted on AWS us-east, so I spin up an EC2 instance in us-east as well as 1 Linode for each of their datacenters and ping each Linode from the EC2 instance. It isn’t an extensive test but should give a good ballpark for anyone considering the same thing:

  • Newark: 9ms
  • Atlanta: 16-19ms
  • Dallas: 34-35ms
  • Fremont: 72ms