Simplified Deployments of Distributed Data Architectures

Project Details

Project Lead
Peter Klipfel 
Project Manager
Peter Klipfel 
Supporting Experts
Gregor von Laszewski  
University of Colorado at Boulder, Computer Science  
Computer Science (401) 


Applications that require elastic, distributed data architectures have large overhead in setup and deployment. I would like to automate deployments and write an opinionated framework for distributed data applications. This framework will provide solutions to schema disparity and codebase divergence. Current methods of deployment (Chef, puppet, etc.) are not set up to deal with elastic clouds; however there are newer, younger frameworks that make this easier (maas, docker, juju, etc.) Testing this framework requires clouds to deploy to.

Intellectual Merit

This project will lower the bar to entry for people that are new to the field, allowing for further innovation. It will also allow those with more skill in the subject to spend more time building novel applications rather than doing mundane configurations. The automation of deployments also decreases the likelyhood of human error and spent time debugging.

Broader Impacts

This research has the potential to help industry deliver more advanced cloud applications to users. The social impact of that is only bounded by the creativity of people to come up with cloud applications. It also has the potential to bring more people into the computing space as the bar for deploying scalable applications lowers.

Scale of Use

I am going to be running deployments repeatedly. I do not need powerful machines. I basically need a bunch of amazon micro instances (Amazon's 750 free hours isn't enough because I am frequently spinning nodes up and down). I would like to use around 10 small servers for testing the architecture. It might be good to have some slightly larger machines to test deployments of heavier applications.