The NMS group at MIT CSAIL conducts broad research into networked and mobile systems. Current topics include Internet architecture and protocols, mobile and sensor computing, datacenter networks, learning-based network systems, systems and networks for machine learning, optical reconfigurable networks, wireless networks, peer-to-peer networks, LLMs for systems. The group was formed in 1998. Previous notable projects include Chord, XCP, RON, r-BGP, Copa, Pensieve, Homa, CCP, CausalSim, SiP-ML, ABC, TopoOpt, Lightning, Trio-ML.
Research areas
Internet architecture & protocols
Designing the protocols and architectural principles that hold the Internet together.
Datacenter networks
Congestion control, transport, and topologies for high-performance datacenters.
Agentic AI for System Optimization
Building agentic AI that automates the design and optimization of computer systems.
Systems & networks for machine learning
Infrastructure that makes large-scale ML training and inference fast.
Learning-based network systems
Using machine learning to optimize and adapt networked systems.
Optical reconfigurable networks
Reconfigurable optical fabrics for AI clusters and datacenters.
Wireless networks
Rate control, sensing, and protocols for modern wireless systems.
Mobile & sensor computing
Systems that sense, compute, and communicate on mobile and embedded devices.
Peer-to-peer networks
Decentralized lookup, routing, and distributed system design.
Recent publications
All publications →Life in the NMS group
Group retreat