Part of the problem is that EC2 is such a whacky environment. Notice Erlang's median connection time was tiny but its average had huge outliers.
Read the timing numbers with a reasonable portion of salt. It's EC2 we're talking about here.
Read the timing numbers with a reasonable portion of salt. It's EC2 we're talking about here.
(I know it's then close to 'dedicated' hosting, but people running benchmarks also want the quick setup and discard of virtualized instances. This would be a hybrid offering that ensures their cloud services is always chosen for such benchmarking comparisons. Of course this offering would not be good for cross-cloud comparisons, because it's not representative of their usual offerings.)
That's called EC2 cluster compute.
- Using VPC you can specify that your instance should run on dedicated hardware [1]
- Cluster Compute instances [2] most likely (see [3]) run on dedicated hardware too
[1] http://aws.amazon.com/dedicated-instances/
[2] http://aws.amazon.com/ec2/#instance
[3] https://forums.aws.amazon.com/message.jspa?messageID=238197#...
If you were to conduct a test like this, where would you go for more reliable performance from hardware?