Timothy Wood Research Group

George Washington University

How can virtualization and new network programmability improve cloud performance and reliability? The GW Cloud Systems Lab develops systems-level tools that use virtualization and programmable network elements to improve performance, efficiency, and reliability in cloud data centers. Researchers build and evaluate serverless and edge platforms (Mu Serverless, EdgeOS) to tackle autoscaling, placement, and fast startup for edge/cloud functions. The group combines programmable switches, SmartNICs, and hosts to design terabit-scale network monitoring and NFV-aware load balancing. Experimental work uses implementation, benchmarking, and deployment studies to measure performance and isolation tradeoffs in real systems.