Automaton automates the process of executing benchmark applications on various resources (e.g., infrastructure clouds and clusters). Ideally this will include all steps from provisioning cloud resources, downloading, configuring, and compiling the benchmark applications, executing the applications, gathering the results, and generating graphs.

Automaton requires that numpy, fabric, boto, and matplotlib libraries be installed.

The easiest way to install these libraries is through pip.

Numpy can be installed with the command:

pip install numpy

All other libraries can be automatically installed using pip by giving the command:

pip install -r requirements.txt

After installing the required libraries, Automaton should be ready to run. Check by running Automaton with:

python ./automaton.py -h

If a help message is shown, then the required libraries have been installed successfully.

Automaton has a global configuration files that needs to be preconfigured before benchmarks can be run -- etc/global.conf. Because Automaton uses ssh to connect to servers, you must provide the path to your public and private keys. Assuming you already have generated a key pair (if not, see Github's guide to generating ssh keys), the two lines that need to be edited are:

key_path = path/to/public/key
ssh_priv_key = path/to/private/key

In addition to configuring etc/global.conf with your ssh keys, Automaton also requires that your credentials from the cloud server that you are accessing be available.

For example, using the Nimbus cloud on FutureGrid:

ssh <username>@hotel.futuregrid.org
cd .nimbus
cat querytokens.sh

Run the two export commands on the local machine.

Automaton should now be configured and ready to run.

Deploying the Automaton software, executing the benchmarks, and gathering and generating graphs is accomplished through the use of flags that can be seen by running:

python automaton.py -h

Generally, the first command to run, in order to launch the desired clusters specified in the etc/cloud.conf file is:

python automaton.py -l

Then, the software can be deployed with:

python automaton.py -s

The benchmarks can be executed with:

python automaton.py -e

Executing the benchmarks may take some time to complete, but once they are finished running the logs can be gathered with:

python automaton.py -o

And finally the graphs generated with:

python automaton.py -p

A specific instance or all of the running clusters can be terminated with:

python automaton.py -t <all/instance id>

1. Read: http://www.python.org/dev/peps/pep-0008/
2. Run pep8 and make sure your code conforms.