Learning-Driven Network Protocols

​​​

• DOTE: Rethinking (Predictive) WAN Traffic Engineering. NSDI 23. Best Paper Award.

• Verifying Generalization in Deep Learning. CAV 23.

• Verifying Learning-Augmented Systems. SIGCOMM 21.

• Towards Scalable Verification of Deep Reinforcement Learning. FMCAD 21.

• Online Safety Assurance for Learning-Augmented Systems. HotNets 20.

• MPCC: Online Learning Multipath Transport. CoNEXT 20.

• PCC Proteus: Scavenger Transport and Beyond. SIGCOMM 20.

• Robustifying Network Protocols with Adversarial Examples. HotNets 19.

• Verifying Deep-RL-Driven Systems. NetAI 19.

• A Deep Reinforcement Learning Perspective on Internet Congestion Control. ICML 19.

• TeaVaR: Striking the Right Utilization-Availability Balance in WAN Traffic Engineering. SIGCOMM 19.

• PCC Vivace: Online Learning Congestion Control. NSDI 18.

• Congestion Control for Future Mobile Networks. CHANTS 18.

• Network-Model-Based vs. Network-Model-Free Approaches to Internet Congestion Control. HPSR 18.

• Learning to Route. HotNets 17.

• Learning to Route with Deep RL. Deep Reinforcement Learning Symposium, NeurIPS 17.

• PCC: Re-architecting Congestion Control for Consistent High Performance. NSDI 15.

 

 

Next-Generation Congestion Control

​
See accessible overview of congestion control here.
 

• Leveraging eBPF to Make TCP Path-Aware. TNSM (2022).

• MPCC: Online Learning Multipath Transport. CoNEXT 20.

• On the Future of Congestion Control for the Public Internet. HotNets 20.

• PCC Proteus: Scavenger Transport and Beyond. SIGCOMM 20.

• A Deep Reinforcement Learning Perspective on Internet Congestion Control. ICML 19.

• Axiomatizing Congestion Control. SIGMETRICS 19.

• PCC Vivace: Online Learning Congestion Control. NSDI 18.

• Congestion Control for Future Mobile Networks. CHANTS 18.

• Network-Model-Based vs. Network-Model-Free Approaches to Internet Congestion Control. HPSR 18.

• Congestion-Control Throwdown. HotNets 17.

• PCC: Re-architecting Congestion Control for Consistent High Performance. NSDI 15.
   

​

Next-Generation Datacenters

 

• Approximate Moore Graphs are Good Expanders. Journal of Combinatorial Theory (Series B), 2019.

• Beyond Fat-Trees without Antennae, Mirrors, and Disco Balls. SIGCOMM 17.

• Xpander: Towards Optimal-Performance Datacenters. CoNEXT 16 and HotNets 15.

• Explicit Expanding Expanders. Algorithmica, 2017.

 

​

Securing the Internet’s Infrastructure

 

• A Devil of a Time: How Vulnerable is NTP to Malicious Timeservers? NDSS 21.

• DISCO: Sidestepping RPKI’s Deployment Barriers. NDSS 20.

• Perfect is the Enemy of Good: Setting Realistic Goals for BGP Security. HotNets 18.

• Preventing (Network) Time Travel with Chronos. NDSS 18. Awarded the IETF/IRTF Applied Networking Research Prize (2019). 

• Are We There Yet? On RPKI’s Deployment and Security. NDSS 17.

• Rethinking Security for Internet Routing. CACM, 2016.

• Measuring and Mitigating AS-Level Adversaries against Tor. NDSS 16.

• Jumpstarting BGP Security with Path-End Validation. SIGCOMM 16 and HotNets 15. Awarded the IETF/IRTF Applied Networking Research Prize (2017). 

• BGP Security in Partial Deployment: Is the Juice Worth the Squeeze? SIGCOMM 13. Awarded the IETF/IRTF Applied Networking Research Prize (2014). 

• Let the Market Drive Deployment: a Strategy for Transitioning to BGP Security. SIGCOMM 11.

• How Secure are Secure Interdomain Routing Protocols? SIGCOMM 10.

​