In the aftermath of the Heartbleed security vulnerability in April 2014, the Linux Foundation set up the Core Infrastructure Initiative (CII) to help fund and support critical open-source projects. Now more than a year later, CII is funding the Census Project to better identify and score potentially vulnerable open-source projects that could benefit from CII’s support.
To date, CII has been assessing projects in need but has realized that a common method and deep assessment of code bases is required to truly identify the projects most in need, Emily Ratliff, senior director of Infrastructure Security at Linux Foundation, said. To date, CII has helped fund multiple projects, including OpenSSL, and announced $452,000 in funding in June for three open-source projects to help improve code quality.
“What’s new is that the Census Project quickly automates the collection and analysis of data on different open-source projects and ultimately creates a risk score for each project based on the results,” Ratliff told eWEEK. “The goal is that this program will expedite CII’s decision-making around grant proposals, ensuring funds are delivered as fast as possible to the software most in need.”
Ratliff herself only recently took on a leadership role in CII as a dedicated resource helping to guide the effort forward.
Ratliff explained that the new risk score for open-source projects is calculated by the Census program based on a number of parameters, including the number of contributors and previously disclosed Common Vulnerabilities and Exposures (CVE) vulnerability identifiers. The risk score calculation is done as an open-source effort, and Ratliff is hopeful that individuals will submit suggestions and improvements to the heuristics to improve the calculation of the score.
“We will rerun the Census periodically to see how project scores change over time, which will happen in large part due to changes in the contributor count, the project’s popularity and the CVE count,” she said. “We are really hoping that this project will open the dialogue around appropriate metrics for assessing open-source project risk.”
The Census Project does not, however, have a mandate to perform static or dynamic code analysis to examine application code for any vulnerabilities. Ratliff said that CII expects open-source developers to use static and dynamic analysis types of tools and perform analysis themselves.
“If they are doing this type of analysis, then their risk score should be reduced over time,” she said.
There have been multiple efforts over the past decade to help open-source projects conduct static analysis on code. One of those efforts is Coverity Scan, which first got started in 2008 thanks in part to a grant from the U.S. Department of Homeland Security (DHS).
A goal of the Census Project is to help identify open-source projects in need of support, though Ratliff noted that there is no specific funding formula in place that is directly correlated to the risk score.
“All funding will be proposed and approved by individuals based on actual project need,” she said. “The goal of the Census Project is to quickly filter projects to the front of the queue for deeper evaluation by the members and advisory board.”
While CII will help fund projects to fix flaws and improve security, Ratliff said CII does not have any plans in place to reward security researchers for disclosing flaws through some form of bug bounty program. Bug bounty programs have become increasingly popular in recent years with multiple technology vendors, including Yahoo, Google, Mozilla and Facebook.
There is still a lot for CII to do in the months ahead, according to Ratliff.
“We remain focused on our mission to identify and support the world’s most critical open-source infrastructure,” she said. “We will also be establishing a set of best practices for open-source development as it relates to security. Look for details on this in the coming months.”
Sean Michael Kerner is a senior editor at eWEEK and InternetNews.com. Follow him on Twitter @TechJournalist.
