Purdue Polytechnic researchers are examining the structure and culture of open source communities to help them be more productive.
With funding from Red Hat, one of the largest supporters of open source communities, Sabine Brunswicker and a research fellow have conducted studies to help identify what makes them successful and to discover inefficiencies. Brunswicker is an associate professor of innovation in the Department of Technology Leadership and Innovation and director of the Research Center for Open Digital Innovation at Purdue. Red Hat’s funding supports a two-year research fellow in the center.
Their studies have focused on the users of Openstack.org. OpenStack is a set of open source software tools for managing and building public and private cloud computing platforms. It is used by companies and organizations around the world, including Red Hat, Xfinity, CERN, and Volkswagen. Brunswicker said OpenStack’s community -- which includes end users, vendors and corporations – is a perfect setting to understand the modern version of an open source community.
“OpenStack is a big, self-organizing community. Red Hat and other players are trying to understand how they thrive. There is a specific service model around open source software, and it is very important to understand how good it is and how well it is progressing. The OpenStack community is quite flourishing, and it is important to understand why,” Brunswicker said.
To tackle their task, Brunswicker said they have identified three areas to research. The first two areas have been completed, and research on the third area will take place in Spring 2017.
Initially, Brunswicker and her team focused on the networking layer of the community, analyzing millions of lines of code to find collaborative patterns. They also examined automated testing tools used in the community. The results of these studies have contributed to useful knowledge in the academic disciplines of information systems and social science.
In the spring, they will examine the stability of the underlying code over time, predictive modeling of the platform’s technology evolution, and how the technology compares to other computing products.
“Our high-level message is that this is quite different from original open source communities,” Brunswicker says. “We will turn our discoveries into an algorithm or analytics that can support Red Hat’s decision making. This is big data research.”
Brunswicker will offer a half-day network analytics science workshop based on her experiences with this project to campus researchers in Spring 2017.