In this project student(s), will deploy a CloudStack based cloud and use mesos to schedule tasks on the resulting virtual clusters. They will show how to run traditional parallel application with MPI and big data application with Hadoop/Map Reduce.
This is primarily an integration project that will result in a proof of concept. Bootstrap code will most likely be created to dynamically provision the virtual clusters and configure mesos on the resulting nodes. We envision a set of puppet or chef recipes and some Python boostrapping code. Potentially this could be an extension of the apache incubating Provisionr project
Language: Java, bash, python
Knowledge: system administration