Crovella Wins NSF Grant for First-of-Its-Kind Domain Name System Census
Research will analyze DNS configurations and management practices using pioneering measurement strategies
When most people type a URL into a browser, they don’t think about the complex machinations that connect them to the website they want to browse. They hit “enter” and assume the site will pop up as intended.
That depends on the Domain Name System (DNS) functioning like it should. The DNS facilitates the connection between what you type in (a name) and the Internet Protocol address (a number) that produces the desired information.
While the DNS is used universally across countries and continents, no single organization controls it, notes Boston University Professor of Computer Science and Computing & Data Sciences (CDS) Mark Crovella. Anyone or anyplace with an identity (such as BU’s www.bu.edu) can add whatever mappings it wants to the DNS. And that means until now, no single entity has had total knowledge of what the DNS stores.
Crovella and two fellow researchers, University of Wisconsin Professor of Computer Science Paul Barford, and Colgate University Professor of Computer Science Joel Sommers, aim to change that with a new project analyzing configurations and management practices in the DNS that recently earned a $525,000 standard grant from the National Science Foundation (NSF).
“The NSF has funded us to build the most complete and extensive census of the contents of the DNS ever assembled,” says Crovella, also the chair of Academic Affairs for CDS. “Using that census, we will analyze the DNS contents in novel ways (based in part on network analysis) to extract key properties of the DNS contents.”
Because Crovella, Barford (who earned his PhD at BU), and Sommers lack the permission of the millions of DNS users to simply download their contents, they must devise what Crovella terms “ingenious measurement strategies.” He notes one census is roughly a terabyte of data, and “we will perform many such censuses.”
“The NSF has funded us to build the most complete and extensive census of the contents of the DNS ever assembled,” says Crovella, chair of Academic Affairs for CDS. “Using that census, we will analyze the DNS contents in novel ways (based in part on network analysis) to extract key properties of the DNS contents.”
Crovella expects information from the large-scale database the researchers plan to build to be available to the community by the end of the three-year project.
“Because the DNS is so crucial to the proper functioning of the internet, knowing exactly what is actually in the DNS is expected to lead to better detection of malicious activity, better diagnosis of configuration errors, and a more complete understanding of how the DNS affects the everyday performance of the internet,” Crovella says. “We will be applying data mining and machine learning to the results of our DNS census to begin to attack these problems.”
When describing his fellow researchers and long-time collaborators, Crovella says he has put together the “right” team to attack this problem.
“Paul, Joel, and I have complementary strengths, and we are united by our curiosity to uncover a complete picture of the DNS and discover the methods and analyses needed to do that.”
Toni Fitzgerald, CDS Contributor