I'm not an expert too, and I also understand possible "fight for the resources" between VMs. I really thought that testing results would be awful. But, probably because I use so-called 'Compute-optimized' instances (see description here: http://aws.amazon.com/ec2/instance-types/), testing results are pretty stable. And this part of the description gives me a large hope in their overall stability (remember, I'm not an expert, especially in hardware): "Each virtual CPU (vCPU) on C3 instances is a hardware hyper-thread from a 2.8 GHz Intel Xeon E5-2680v2 (Ivy Bridge) processor". On Fri, Mar 28, 2014 at 7:05 PM, Giovanni Mascellani < mascellani@xxxxxxxxxxxxxxxxxxxx> wrote: > Hi. > > Il 28/03/2014 08:23, Artem Iglikov ha scritto: > > Hmm. Today I restarted all workers simultaneously, and they performed > > equally. Probably, first time worker1 was running some unnecessary > > service that I forgot to shutdown. > > I'm not an expert on this, but I understand that EC2 machines are > virtual machines that share their CPU with other virtual machines. Thus, > depending on the load of the "companions" instances, they may perform > better or worse (I would say that Amazon guarantees for a minimum > performance, not for a maximum one). > > But, again, I'm not expert of this field. > > Giovanni. > -- > Giovanni Mascellani <giovanni.mascellani@xxxxxx> > PhD Student - Scuola Normale Superiore, Pisa, Italy > > http://poisson.phc.unipi.it/~mascellani > > -- Artem Iglikov