24 October 2019 – Associate Professor Chan Mun Choon and Computer Science PhD students Nishant Budhdev, Pravein Govindan Kannan and Raj Joshi won the highly coveted Facebook Research award on 25 September this year.
The team’s proposal was one of six research ideas chosen for their significant potential to improve global networking systems, data centres, and edge networks. Recipients of this award will receive an unrestricted gift of US$50,000 to fund their projects. A total of 71 networking systems research proposals from 17 countries were submitted for the Facebook Research award in April this year.
A/P Chan and his PhD students Nishant, Pravein and Raj received the award for their proposal to develop a tool to accurately detect and diagnose bugs and network issues. “Network monitoring and debugging have become more challenging as link speeds now reach 400 gigabit per second and the number of end-points in a network can cross to over 100,000 points. Furthermore, as applications become more diverse and complicated, network interactions have also become more complex,” explained A/P Chan.
One of the challenges network operators face today is the difficulty of locating the root cause of problems in the network. “With the complexity of today’s networks, it is almost impossible to find the root cause of certain problems by studying a network switch’s local information. Operators will need to have fine-grained packet-level and network-wide visibility to do so,” A/P Chan added.
To solve this problem, the team proposed to develop a network-wide synchronised record and play framework to record and play back the network’s traffic conditions. The system will record packet-level activities within the network and provide a network-wide offline replay of the network. This will allow operators to find the root causes of network issues using SQL queries and dashboards. “We hope that our tool will help users to accurately detect and diagnose events like microbursts, misconfigurations, protocol bugs, and more. In addition, we believe that that our system can also be used in regression test suites for large software development and to help the networking research community understand network traffic behaviour in the wild,” said A/P Chan.
“Our work will be the first of its kind system that records and replays the entire network in a scalable manner. Winning this award has validated our idea and provided a major boost to our research in data plane programmable networks,” A/P Chan added.
Nishant, Pravein and Raj are fifth year PhD students with the Department of Computer Science in NUS Computing. Nishant and Pravein are under the supervision of A/P Chan, while Raj is under the supervision of A/P Ben Leong.
Facebook Research, 25 September 2019