When CCIS professors CCIS professors David Choffnes, Alan Mislove and Christo Wilson set out to delve into everyday issues ranging from understanding how Uber surge pricing works to investigating whether websites are as safe as their certificates claim they are, the trio didn’t expect that seven of their papers would be accepted into the Internet Measurement Conference (IMC) 2015 in Tokyo.

“To have seven papers among the three of us is incredibly rare and highlights not just what a strong group we’ve managed to put together here at Northeastern over the past six or seven years, but also how collaborative we are,” Choffnes says.

In collaboration with five Ph.D. candidates and one undergraduate student, Choffnes, Mislove and Wilson submitted a total of 10 papers with overlapping authorship between the three professors. Choffnes had three accepted, Wilson had four and Mislove had five. But this conference isn’t just an opportunity for the faculty members to shine: several of the papers had students listed as lead authors.

“They’re all authors and lead authors on the papers,” Wilson says. “Most of them are going to be presenting the work at the conference.”

The papers

The three professors worked together on a paper titled An End-to-End Measurement of Certificate Revocation in the Web’s PKI, which looks at whether security certificates are rejected when they should be – they aren’t. Secure Sockets Layer (SSL) technology sets up encrypted links between servers and browsers to keep user information safe. Servers using this technology are issued certificates that browsers must validate to make sure that users aren’t making themselves vulnerable to malicious attacks. And while browsers should be alerted if there’s been a security breach on the server, “they’re not doing a particularly good job of letting people know,” Wilson says.

In Peeking Beneath the Hood of Uber, Mislove and Wilson tried to understand how everyone’s least favorite Uber feature – surge pricing – works, and whether it actually works the way Uber wants it to. Short story: probably not, unless Uber’s goal is to reduce demand. Uber shows users the closest eight drivers, so the professors blanketed New York City, placing smartphones with Uber accounts roughly 650 ft. apart in order to see all the city’s drivers. What they found was interesting: because of how Uber divides cities into small areas, surge prices in one neighborhood don’t necessarily mean that the adjacent area will experience the same surge pricing. So next time the app shows a 1.8 surge charge, walk a few blocks and try again – you might get lucky. They also found that it’s not just users who are frustrated by surge pricing. Drivers aren’t fans either, because they’re left without customers. “What the surge really does is just kill demand,” Wilson says.

If you’ve ever wondered why Google search results sometimes differ based on where you’re searching from, Mislove and Wilson’s Location Location Location: The Impact of Geolocation on Web Search Personalization has that answer – your location is a factor in how Google personalizes search results. Mislove and Wilson found this out by creating a fake browser and using Google Search from fake locations at three levels: Cleveland by voting district, Ohio by county and the United States by state. After about half a mile, results start to differ – which means that Google might offer you different dinner options in Mission Hill to your friend in Allston, or even that Google might tailor your search results on presidential candidates based on the political leanings of your neighborhood. “The only time in which we could see two different locations getting the same result was when they were extremely close together,” Mislove says.

Wilson also co-authored Offline Downloading in China: A Comparative Study, a paper with researchers in China, where in order to combat bandwidth issues, they developed a technology to help users decide how best to access files: stream, download from the website, or download via peer-to-peer sharing.

Mislove, currently on sabbatical in one of the world’s happiest nations (Denmark), worked with colleagues in Copenhagen on Opportunities and Challenges in Crowdsourced Wardriving. The paper delves into how companies like Google and Apple compile Wifi localization databases, or lists of Wifi networks that tell servers where a smartphone is when users turn to apps like Google Maps or Apple Maps for directions. Because using GPS signals wastes time and drains battery life, smartphones use Wifi access points to figure out where they are. Databases that major companies maintain can include billions of access points. “So every time you ask Google, it cross references this database and says, ‘Hey, here’s the most likely location,’” Mislove says. In Copenhagen, researches handed smartphones to 1,000 students and collected data from those phones to understand the process of creating these databases.

Mislove and Choffnes also collaborated on Identifying Traffic Differentiation in Mobile Networks. Through their research for that paper, the two found that the Federal Communications Commission (FCC) laws that upheld net neutrality and prohibited mobile networks from giving different treatment to different applications were being obeyed. Before the FCC ruling, the two had been working on a hunch that traffic differentiation would probably be an issue in a mobile environment where bandwidth is limited. AT&T, Sprint, T-Mobile and Verizon passed the test – they were not interfering with the rate at which data could be sent over their networks, so customers of those networks could continue to stream high quality YouTube videos from their phones without worries. Customers of smaller networks weren’t so lucky, and Sprint users also had to contend with images being transcoded, or downloaded at a low quality. “Our original tests were before the new FCC rules,” Choffnes says. “When we tested after the FCC rules took effect, we found that those cases where there was shaping, there is no longer shaping in those networks. So it seemed like they did change their policies in response to the rules.”

Choffnes’ third paper, Investigating Interdomain Routing Policies in the Wild, was in collaboration with researchers from Stony Brook University, University of Southern California and Universidade Federal de Minas Gerais in Brazil. The paper describes techniques to reveal the information that helps form the structure of the Internet, which can help in determining important properties such as Internet reliability, stability and security.

With the seven papers that Choffnes, Mislove and Wilson had accepted into the IMC conference, they made up just over 15 percent of the conference program – not an easy feat to top, but they hope future students will rise to the challenge.

“This will be a tough act to follow,” Choffnes says. “The first thing we told the students after hearing that we had seven papers is that we’ll have to have eight next year. Several students had very panicked looks on their faces before they realized we were kidding.”