As I get closer to graduation, I've been looking at possibilities of starting a company and running those resources using private resources. There are two ways of developing a private cluster: (1) by setting up your own environment locally, (2) by taking advantage of cloud infrastructure environments.
Most computational research is dependent upon large HPC environments. This section is really a list of notes on how to do some common administrative tasks. Most of the information here is really some architectural notes for the implementation of future architectural capabilities of PYPOSPACK.
Mostly beause AWS is currently the biggest player in the space. And for purposes of self-preservation, it is always beneficial to always develop skills in the technology which is greatest demand.
So this isn't really necessary, but the AWS documentation is somewhat difficult to navigate.
1. Provision an EC2 host. Since a webserver is going to up all the time..
2. Change your DNS records. The easiest thing to do is to add a CNAME record which points to the hostname of your EC2 instance.
I'm interested in AWS clusters because it appears to be the most flexible method of getting on-demand computing resources. More information here.