I wish the issue was explained a bit more in the article. In my opinion if the blocking was due to host-names then it had nothing to do with EC2. If the blocking was due to re-use of IP addresses then maybe it had something to do with EC2 (but the article didn't explicitly claim that). As far as I can tell the effect was due to using a outsider server at all (instead of carrying one into the client's own private data center), which I think has nothing to do with having used EC2.